The Association of Vitamin D and Its Pathway Genes’ Polymorphisms with Hypertensive Disorders of Pregnancy: A Prospective Cohort Study

Objective: We aimed to explore the effect of single nucleotide polymorphism (SNP) in the genes of the vitamin D (VitD) metabolic pathway and its interaction with VitD level during pregnancy on the development of hypertensive disorders of pregnancy (HDP). Methods: The study was conducted in the Zhoushan Maternal and Child Health Care Hospital, China, from August 2011 to May 2018. The SNPs in VitD metabolic pathway-related genes were genotyped. Plasma 25-hydroxyvitamin vitamin D (25(OH)D) levels was measured at first (T1), second (T2), and third (T3) trimesters. The information of systolic blood pressure (SBP) and diastolic blood pressure (DBP), and the diagnosis of HDP were extracted from the electronic medical record system. Multivariable linear and logistic regression models and crossover analysis were applied. Results: The prospective cohort study included 3699 pregnant women, of which 105 (2.85%) were diagnosed with HDP. After adjusting for potential confounders, VitD deficiency at T2, as well as the change of 25(OH)D level between T1 and T2, were negatively associated with DBP at T2 and T3, but not HDP. Polymorphisms in CYP24A1, GC, and LRP2 genes were associated with blood pressure and HDP. In addition, VitD interacted with CYP24A1, GC, and VDR genes’ polymorphisms on blood pressure. Furthermore, participants with polymorphisms in CYP24A1-rs2248137, LRP2-rs2389557, and LRP2-rs4667591 and who had VitD deficiency at T2 showed an increased risk of HDP. Conclusions: The individual and interactive association between VitD deficiency during pregnancy and SNPs in the genes of the VitD metabolic pathway on blood pressure and HDP were identified.


Introduction
Hypertensive disorders of pregnancy (HDP), including gestational hypertension, preeclampsia, eclampsia, pregnancy complicated with chronic hypertension, and chronic hypertension complicated with preeclampsia [1], accounted for nearly 18% of all maternal deaths worldwide [2]. Its increasing prevalence and related risks for maternal and child health as well as cardiovascular diseases later in life has garnered great attention in the field of public health [3,4]. The risk factors for HDP are advanced age, primipara, multiple pregnancy, family history of hypertension, high pre-pregnancy body mass index (BMI), and high basal blood pressure [5].
Approximately 5% to 7% of pregnancies are complicated by preeclampsia [6]. While the cause of preeclampsia is not fully discerned, previous studies have suggested that abnormal placentation and angiogenesis were central to the pathogenesis of this syndrome [6]. In recent years, growing evidence of the association between maternal hypovitaminosis D and increased risk of HDP has been suggested [7,8]. Compared to non-pregnant state, there are significant changes in vitamin D (VitD) metabolism during pregnancy, and the serum levels of VitD binding protein (VDBP) [9], as well as the active form, 1,25-dihydroxyvitamin (1,25(OH) 2 D) [10], increased notably. It is believed that not only the kidneys but also the placenta and decidua produce and secret 1,25(OH) 2 D during pregnancy [11]. Moreover, VitD receptors and related metabolic enzymes have been discovered in the placenta and decidua [12], indicating a potential role for VitD in implantation and placental function, outside of its well-established role in skeletal health [13].
To date, trial evidence appears insufficient to lean towards a protective effect of VitD supplementation during pregnancy against the risk of preeclampsia owing to small sample size or low study quality [14,15]. In addition, findings from observational studies in regard to the association between maternal VitD status and HDP are discrepant due to the large heterogeneity between study designs, lack of adherence to standardized outcome definitions, and different gestational weeks of VitD detection [8,16]. On the other hand, genetic variants in the VitD metabolic pathway have also been shown to participate in the pathogenesis of blood pressure increase and preeclampsia [8,17], which suggests a possible interaction between VitD and its pathway gene variants for HDP. The concentration or effect of VitD can be highly regulated due to the variation of key protein expression or activity. 25(OH)D is the main circulating metabolism and is considered the biological marker of VitD status. The main metabolic enzymes involved in the synthesis, transport, reabsorption, and inactivation of VitD include 25-hydroxylase (CYP3A4), 1-hydroxylase (CYP27B1), vitamin D-binding protein (GC), 24-hydroxylas and metaling (LRP2), and 24-hydroxylase (CYP24A1). Moreover, VitD receptor (VDR) regulates VitD metabolism through binding 1,25(OH) 2 D [18].
So far, most studies have only focused on the relationship between VitD status during pregnancy or gene variation in the VitD metabolic pathway and HDP, without considering the possible interaction between them. This study aimed to explore the association of VitD status in three trimesters of pregnancy with the risk of HDP, and to explore the interactive effect between maternal VitD level and genetic variants in the VitD metabolic pathways (GC, CYP24A1, CYP3A4, CYP27B1, LRP2, VDR) on gestational blood pressure and HDP.

Study Design and Participants
The Zhoushan Pregnant Women Cohort (ZPWC) is an ongoing prospective cohort, conducted in Zhoushan Maternal and Child Health Care Hospital, Zhejiang. This study was based on the data of ZPWC from August 2011 to May 2018. We recruited pregnant women aged between 18 and 45 years at their first prenatal visit. A more detailed description of the inclusion and exclusion criteria can be seen in a previous study [19]. In addition, pregnant women without extreme/missing information of blood pressure and who measured plasma 25(OH)D levels in the first, second, or third trimester were included in the study. In addition, because gestational hypertension (GH), preeclampsia, and eclampsia are different from pregnancy complicated with chronic hypertension and chronic hypertension complicated with preeclampsia in pathogenesis and clinical treatment, pregnant women with chronic hypertension before pregnancy were also excluded [1]. Informed consent was obtained from all participants before the investigation. The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board of Zhoushan Maternal and Child Health Care Hospital on 9 January 2011 (Ethical Approval Code: 2011-05).

Collection of Data and Blood Sample
The interviewers conducted face-to-face questionnaire surveys. Socio-demographic characteristics, lifestyle, and health behavior in the first (T1: 8th-14th gestational week), second (T2: 24th-28th gestational week), and third (T3: 32nd-36th gestational week) trimester, and 42nd day postpartum were collected. At each visit, professional nurses and inspectors were responsible for drawing and centrifuging fasting venous blood samples under 4 • C and separating the plasma and white blood cells, which were then stored at −80 • C until use.

Covariates Assessment
According to Endocrine Society Clinical Practice Guidelines, we defined plasma 25(OH)D < 20 ng/mL (50 nmol/L) as VitD deficiency [20], and 25(OH)D concentrations ≥ 20 ng/mL as VitD non-deficiency. The change of 25(OH)D level during pregnancy is defined as a difference of 25(OH)D level between three trimesters. The following parameters were also defined: Pre-pregnancy body mass index (BMI) = weight (kg)/height 2 (m 2 ), gestational weight gain (continuous) = the weight on the day of VitD test at T1, T2, or T3, the pre-pregnancy weight, educational level (senior high school and below, college and above), gravity (1, ≥2, missing), parity (0, ≥1, missing), basal blood pressure (the level of blood pressure at the first prenatal examination or early pregnancy, continuous), the seasons of blood pressure measurement (divided as followed: spring (March to May), summer (June to August), fall (September to November), and winter (December to February) based on the sunshine intensity and duration in different months [21]).

HDP Definition
In perinatal care, SBP and DBP would be routinely measured [22]; we extracted the data from EMRS. In a sitting position, blood pressure measurement was performed from the right hand with a standard mercury sphygmomanometer. GH onset was defined as SBP ≥ 140 and/or DBP ≥ 90 mm Hg after the 20th gestational week (according to last menstruation date and B-ultrasound) in at least two consecutive examinations [23]. On the basis of GH, urinary protein ≥ +1 on a dipstick was defined as preeclampsia [1]. Eclampsia was defined as the presence of new-onset grand mal seizures in a woman with preeclampsia [24]. GH, preeclampsia, and eclampsia were combined as the group of HDP in later analysis.
The conventional phenol-chloroform extraction method was used to extract DNA from the peripheral blood leukocytes, which was then stored in TE-buffer at −80 • C. DNA was diluted to 10 ng/µL using a Nanodrop ® ND-1000 Spectrophotometer (Thermo Fisher Scientific Inc., Wilmington, NC, USA) for SNP analysis. A Sequenom MassARRAY iPLEX Gold platform (Sequenom, San Diego, CA, USA) was used for SNP genotyping. The call rate of these SNPs was over 98%, which conformed to the Hardy-Weinberg equilibrium.

Statistical Analysis
The characteristics between HDP and non-HDP groups were compared by t-test for continuous variables and by chi-squared test for categorical variables. Latent mixture modeling (PROC TRAJ) was used to identify subgroups that shared similar VitD patterns. Model fit was assessed using the Bayesian Information Criterion. We initiated a model with three trajectories, and then compared the BIC to that with two. The model with three trajectories identified fit best [27] ( Figure S1). Restricted cubic spline (RCS) analyses were used to characterize the dose-response association and explore the potential linear or nonlinear relationship of 25(OH)D level in three trimesters, the change of 25(OH)D level during pregnancy with blood pressure in three trimesters, and HDP. Multivariable adjusted analyses with three knots were used. Test result for nonlinearity was checked first. If the test for nonlinearity was not significant, test result for overall association and linearity was checked, with a significant result indicating a linear association [28]. Multivariate adjusted RCS analysis showed that there was no nonlinear association of 25(OH)D level in three trimesters, the change of 25(OH)D level during pregnancy with blood pressure, and HDP during pregnancy (P for non-linear > 0.05) ( Figures S2-S6). The Hardy-Weinberg equilibrium (HWE) of genotyped SNPs was tested using the χ 2 test.
A multiple linear regression model and a multivariate logistic regression model, combined with a crossover analysis method were utilized to explore the association between VitD and its metabolic pathway-related gene variants as well as their interactions with SBP, DBP, and HDP. The generalized linear model was used to analyze the relationship of the change of 25(OH)D level during pregnancy with SBP and DBP, and the multivariate logistic regression model was used to analyze the association between the changes in 25(OH)D levels and the trajectory of VitD during pregnancy with HDP. Models were adjusted for the following potential confounders: pre-pregnancy BMI, maternal age, gestational weight gain, gestational week, educational level, parity, basal blood pressure, and the seasons of blood pressure measurement.
β (se) for linear regression, ORs, and corresponding 95% CIs for logistic regression were calculated, respectively. All test results were considered statistically significant at a value of p < 0.05. RCS analyses were performed using R software (version 3.6.3); the other analyses were performed using SAS (version 9.4, SAS Institute, Cary, NC, USA).

Subject Characteristics
The demographic characteristics of participants with HDP or non-HDP were compared and are shown in Table 1. The prospective cohort study included 3699 pregnant women, of which 105 (2.85%) were diagnosed with HDP. The mean age was 29.30 ± 3.95 years for HDP participants and 28.67 ± 3.64 years for non-HDP participants. Compared with non-HDP participants, HDP women had higher pre-pregnancy BMI (21.16 ± 2.91 kg/m 2 vs. 23.62 ± 4.05 kg/m 2 , p < 0.0001). The SBP and DBP levels in three trimesters were higher in HDP than non-HDP. However, VitD deficiency in three trimesters, educational level, gravity, and parity were not significantly different between the two groups. The characteristics of participants in the SNP analysis are shown in Table 1. Pregnant women with HDP had higher pre-pregnancy BMI than the non-HDP group. There was no significant difference in weight gain and 25(OH)D level in three trimesters, educational level, gravity, and parity between the two groups (Table S1).

The Association between SNP and HDP
The association of each SNP genotype with SBP and DBP at T1, T2, and T3 are shown in Tables S7-S9, respectively. Polymorphisms in CYP24A1-rs2248137 was significantly associated with higher SBP at T1 and DBP at T2 and T3. Polymorphisms in CYP24A1-rs2762934 were significantly associated with higher DBP at T1 and SBP at T2. Polymorphisms in LRP2-rs4667591 were significantly associated with higher SBP at T1 and DBP at T3. Polymorphisms in GC-rs2070741, rs222020, and rs2298849 were associated with higher SBP at T2. Polymorphisms in LRP2-rs2544390 were associated with higher DBP at T3. Furthermore, polymorphisms in CYP24A1-rs2248137, CYP24A1-rs2762934, CYP24A1-rs6127118, and GC-rs2070741 were associated a higher risk of HDP (Table 5). However, there was no significant association between other genes' polymorphisms and HDP.

The Interaction between Single SNP and VitD Deficiency in Three Trimesters on the Risk of HDP
Results of the crossover analysis are shown in Tables S10-S13. Polymorphisms of seven SNPs (rs16846876, rs2282679, rs17467825, rs2298849, rs2298850, rs3755967, and rs4588) in GC gene and VitD deficiency at T2 might exert interactions on DBP at T2. In addition, VDR-rs2228570 and VitD deficiency at T2 might exert interaction on SBP at T2. Furthermore, women with mutations in CYP24A1-rs2248137, LRP2-rs2389557, and LRP2-rs4667591 and had VitD deficiency at T2 showed increased risk of HDP (Table 6).

Discussion
In the present study, 25(OH)D level at T2, as well as 25(OH)D change between T1 and T2, were significantly inversely associated with DBP at T2 and T3. However, significant associations between maternal VitD deficiency in any trimesters and HDP were not observed. Polymorphism in CYP24A1, GC, and LRP2 was associated with blood pressure, and polymorphism in CYP24A1 and GC was associated with increased risk of HDP. Furthermore, interactive effects between VitD deficiency and polymorphisms in CYP24A1, GC, and VDR genes on blood pressure were identified. Women with polymorphisms in CYP24A1 and LRP2 genes and had VitD deficiency at T2 showed a higher risk of HDP.
Previous findings on the association between VitD level during pregnancy and HDP were not consistent. A prospective observational study conducted in southern China found that there were no significant differences in the risk of HDP among women with different levels of VitD at 16-20-week gestation [29]. A case-control study conducted in Iran found that pregnant women with VitD deficiency (25(OH)D < 20 ng/mL) had higher blood pressure and increased risk of preeclampsia than those with VitD insufficiency (25(OH)D: 20~30 ng/mL) [8]. The prospective Swedish GraviD cohort study, including 1413 pregnant women, found that 25(OH)D was positively associated with T1 blood pressure [16]; however, both 25(OH)D level at T3 and change in 25(OH)D level from T1 to T3 were significantly and negatively associated with preeclampsia, but not with the risk of GH [30]. Another nested case-control study carried out among Australian pregnant women found that higher levels of VitD (25(OH)D > 75 nmol/L) in early pregnancy (10-14 weeks) could prevent the occurrence of early-onset preeclampsia (p = 0.09); however, women with low levels of 25(OH)D (<37.5 ng/mL) in the first trimester of pregnancy had a tendency toward reduced risk of preeclampsia (p = 0.07) [31]. Conflicting data for an association of VitD during pregnancy with HDP results from a number of sources, including large heterogeneity between study designs, different ethnicities, different subtypes of HDP included in the analysis, variable quality of measurement for 25(OH)D, and inconsistent definition of VitD status [32]. On the other hand, studies have shown that the gene variation of key enzymes in VitD synthesis, transport and metabolism pathway would also affect the levels and effects of 25(OH)D and 1,25(OH) 2 D [25,33]. Furthermore, genetic mutations in the VitD metabolic pathway were also associated with increased risk of HDP [8].
The active form of VitD (1,25(OH) 2 D) needs to bind to VDR to exert its biological function. Relevant studies related to genetic variants in the VitD metabolic pathway with HDP were mainly focus on three SNPs (rs2228570, rs731236, and rs1544410) of VDR gene. Rezavand et al. [8] found that, compared with VDR-rs2228570 TC and TT + TC genotypes, the SBP and DBP of CC genotype were higher, and the risk of preeclampsia increased by 1.72 times. However, no association was found between VDR-rs731236, VDR-rs1544410, and preeclampsia. Knabl et al. [34] also reported that there was a strong association between the polymorphisms in rs10735810 and rs1544410 of VDR and the risk of GH. The polymorphisms in rs10735810 affect plasma renin activity and may be associated with a reduced risk of GH [34]. In this study, VitD deficiency at T2 interacted with the variants of VDR-rs2238136 on DBP and VDR-rs2228570 on SBP at T2.
The CYP24A1 gene is located in 20q13-2, which is mainly expressed in the kidney and encodes the catabolic enzymes of 1,25(OH) 2 D and 25(OH)D [35]. Evidence relating to the association between CYP24A1 gene polymorphism and susceptibility to hypertension, especially among pregnant women, is scare. A case-control study among the Chinese Han population found that CYP24A1-rs56229249 significantly decreased the hypertension risk in homozygote and recessive models [36]. In addition, rs2762940 was related to hypertension risk in men, and rs56229249 was a protective factor against hypertension in women [36]. The comprehensive genetic association study in the Women's Genome Health Study (WGHS) found that CYP24A1-rs2296241 showed significant associations with SBP, DBP, mean arterial pressure, and pulse pressure [37]. In this study, we found that gene variants in CYP24A1-rs2248137, CYP24A1-rs2762934, and CYP24A1-rs6127118 were associated with increased risk of HDP. Furthermore, CYP24A1-rs6013897 interacted with VitD deficiency at T2 on HDP. On the other hand, LRP2 is located on 2q24-q31, which is a member of the low-density lipoprotein receptor family and encodes megalin protein.
In the kidney, megalin and cubilin combine together with hydroxylate 25(OH)D 3 into 1,25(OH) 2 D 3 [38]. Studies regarding the association between LRP2 genes and VitD with the risk for HDP are still lacking. This study found that the mutations of LRP2-rs2389557 and LRP2-rs4667591 and VitD deficiency at T2 had a combined effect on the risk of HDP.
The GC gene encodes VitD binding protein (VDBP) [39], which is the major transporter of VitD. About 85% to 90% of 25(OH)D is bound to VDBP in circulation [40]. VDBP can aggravate or enhance various biological processes during pregnancy, such as immune regulation, glucose metabolism, and blood pressure regulation [39]. The GC-1 subtype was more common in pregnant women with preeclampsia than in those without preeclampsia, which was considered as a potential early detection genetic marker for women at risk of preeclampsia [41]. In HIV endemic areas of South Africa, compared with women with normal blood pressure, two SNPs of GC gene (rs4588 and rs7041) were more common in pregnant women with preeclampsia, and were not related to HIV status [42]. Furthermore, GC-rs4588 polymorphism was associated with early-onset (<34 weeks) and late-onset (≥34 weeks of pregnancy) preeclampsia, while GC-rs7041 was associated with early-onset eclampsia [42]. A nested case-control study of 170 American women from Massachusetts tracked the levels of VDBP and 25(OH)D throughout pregnancy to examine whether these biomarkers were associated with blood pressure or the risk of preeclampsia, but found no significant correlation of VDBP or 25(OH)D levels with preeclampsia [43]. At present, the combined effect of GC gene polymorphism and VitD during pregnancy on HDP is not clear. A study focused on preterm birth found that rs7041 variants interacted with VitD at T2 on the gestational week of delivery and preterm birth [44]. Our study found that the variant of GC-rs2070741 was associated with higher SBP at T2 and increased risk of HDP. Mutations at GC rs16846876, rs2282679, rs17467825, rs2298849, rs2298850, rs3755967, and rs4588 interacted with VitD deficiency at T2 on higher DBP at T2.
To our knowledge, this is the first prospective cohort study exploring the association between VitD in three trimesters and VitD pathway gene variants as well as their interactions on SBP, DBP, and the risk of HDP. However, limitations could not be neglected. First of all, some subjects had a lack of 25(OH)D data at T2 and T3, and therefore selection bias might exist. However, subgroup analysis of pregnant women with VitD detected at T1 and T2 showed that the results were almost consistent with the results in the whole study population. Secondly, as the prevalence of HDP in this study was relatively low (2.84%), the association between VitD and different HDP subtypes (GH, preeclampsia, eclampsia) could not be explored. However, studies have shown that, although these subtypes can appear alone, they are progressive manifestations of a single process and share common etiology [45,46]. Lastly, the relatively single ethnic population of this study may also limit the extrapolation of findings.

Conclusions
This study found that the level of 25(OH)D at T1 and T2 was negatively correlated with DBP at T2. In addition, polymorphisms in VitD metabolic pathway genes, including CYP24A1 and GC, increased the risk of HDP. Furthermore, gene variants in CYP24A1 and LRP2 and VitD deficiency at T2 showed combined effect on the risk of HDP, but the specific mechanism remains to be further investigated. The results of this study provide a scientific basis for the clinical detection of VitD during pregnancy and the supplementation of VitD during pregnancy.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/nu14112355/s1, Figure S1. Trajectory of 25(OH)D level during pregnancy. Figure S2. Dose-response relationships of 25(OH)D levels at T1 and T2 with blood pressure at T1 and T2. Figure S3. Dose-response relationships of 25(OH)D levels at T1, T2, and T3 with blood pressure at T3. Figure S4. Dose-response relationships of 25(OH)D levels at T1, T2, and T3 with HDP. Figure S5. Dose-response relationships of the change of 25(OH)D levels during pregnancy with blood pressure at T2 and T3. Figure S6. Dose-response relationships between the change of 25(OH)D levels during pregnancy and HDP. Table S1. Baseline characteristics of pregnant women in SNP analysis. Table S2. The relationship of VitD deficiency in three trimesters with blood pressure. Table S3. The association between 25(OH)D levels in three trimesters and HDP. Table S4. The relationship between VitD deficiency in three trimesters with HDP. Table S5. The association between the change of 25(OH)D levels during pregnancy and HDP. Table S6. The association between the trajectory of VitD during pregnancy and HDP. Table S7. The association of single SNP with SBP and DBP at T1. Table S8. The association of single SNP with SBP and DBP at T2. Table S9. The association of single SNP with SBP and DBP at T3. Table S10. The association of single SNP and VitD at T1 with blood pressure at T1. Table S11. The association between single SNP and VitD at T2 with blood pressure at T2. Table S12. The association between single SNP and VitD at T3 with blood pressure at T3. Table S13. The association between single SNP and VitD at T2 with blood pressure at T3. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available because they contain information that could compromise the privacy of research participants.