全文

Abstract

Objective

The active form of vitamin D (1,25[OH]2D3) has been established to have potent anti-proliferative, immuno-modulatory, and anti-microbial action in addition to its effects on bone. The nuclear vitamin D receptor (VDR) is expressed in the placenta-decidua, regulating genes associated with implantation and implantation immuno-tolerance. If VDR polymorphisms regulate VDR functionality at the placenta-decidua interface, VDR genotypes may be involved in idiopathic preterm birth (PTB).

 

Study design

Maternal and fetal (umbilical cord) blood samples from 33 Jewish and Arab mothers with PTB of a singleton neonate were compared to 98 samples from Jewish and Arab maternal and fetal blood samples from full-term, uncomplicated singleton births. Maternal age and ethnicity were comparable between groups. PCR amplification/digestion identified the VDR SNPs: FokI, ApaI, TaqI, and BsmI.

 

Results

Allele frequency for the FokI VDR in maternal blood samples from preterm births (but not umbilical cord samples) was significantly different (p = 0.01) from that in maternal and umbilical cord blood samples from full-term singleton births, with an odds ratio for FokI carriers of 3.317 (95% CI, 1.143, 9.627) for preterm birth. The FokI VDR variant may therefore be a maternal risk trait for PTB among these women.

 

Conclusion

This study may support a future platform for the study of vitamin D during pregnancy and treatment of selective target populations with vitamin D and/or VDR “tissue-specific therapeutic intervention” for prevention of PTB.

 

Keywords

Placenta-decidua; Preterm birth; VDR polymorphisms; Vitamin D; Vitamin D receptor (VDR)
 

Introduction

Preterm birth (PTB) occurs in approximately 12% of all pregnancies in North America, an incidence which has remained relatively unchanged in recent years and which is only slightly reduced when good antenatal care is consistent [1]. Spontaneous PTB is associated with idiopathic preterm labor or with preterm rupture of membranes (PROM). The incidence of idiopathic PTB is unevenly distributed among fertile women and differs by geographic area, race, and ethnicity [2]. PTB is associated with approximately 70% of all neonatal deaths and more than 75% of early severe neurological, respiratory, and gastrointestinal neonatal morbidity [2]. In the long-term, children born prematurely have an increased risk of cardiovascular disease, hypertension, and diabetes as adults and possibly an increased risk of cancer [3] and [4]. Beyond the associated significant morbidity and mortality for the infants, PTB is also a considerable emotional and financial burden for the family, and involves substantial use of scarce health care resources [5].

 

The etiology of spontaneous PTB is still unknown. PTB is thought to be a multi-factorial event, with potential interplay of various endogenous pathophysiological features such as stress [6], genetic factors such as predisposition to inflammation, with or without evidence of infections [7] and [8], and environmental or occupational factors [9] that may exist singly or appear in combination, to name just a few [10] and [11]. It is unclear, however, whether “the PTB syndrome” is a physiological process comparable to term labor but activated prematurely, or whether it is a uniquely pathological process because of abnormal signaling. The observation that PTB is a repetitive event in some women and has a familial cluster pattern speaks to a possible predisposing genetic basis [2], [12], [13], [14] and [15].

 

The active form of vitamin D (1,25 dihydroxyvitamin D3; 1,25[OH]2D3) has well-established effects on bone metabolism and mineral homeostasis [16]. Data are emerging that maternal vitamin D status in pregnancy may affect intrauterine skeletal mineralization and growth [17]. Recently, however, it has also become clear that 1,25[OH]2D3 has potent anti-proliferative and immuno-modulatory actions [18] and [19] that are not overtly linked to its role as a skeletal regulator. Both the nuclear receptor for 1,25[OH]2D3 (vitamin D receptor; VDR) and the vitamin D activating enzyme 1-α hydroxylase are expressed in a variety of non-musculoskeletal tissues, including the placenta [20]. This raises the possibility that locally generated 1,25[OH]2D3 may be involved in fetal-placental development and function by virtue of regulating cell proliferation and differentiation [21] and/or in the “cross-talk” of signaling growth factors between placenta and fetus [22]. The immuno-modulatory effects of 1,25[OH]2D3 may underscore a role for the vitamin D system in immuno-tolerance at implantation wherein vitamin D status may be involved in early pregnancy complications such as preeclampsia [23]. It has been proposed that (epigenetic) uncoupling of feedback regulation of renal hydroxylases involved in activation of vitamin D during pregnancy may determine vitamin D bioavailability at the feto-maternal interface [24].

 

Central to the functional unit of the vitamin D system is the vitamin D receptor (VDR) that exerts diverse effects including those of bone mineral homeostasis, detoxification of compounds, cancer prevention, and also mammalian hair cycling [16], [25] and [26]. There are four well-characterized di-allelic polymorphisms of the VDR that have been described: FokI C > T (rs10735810) and TaqI T > C (rs10735810) polymorphic sites on the coding sequence, and BsmI A > G (rs1544410) and ApaI G > T (rs7975232) on the last intron. The latter three have been shown to be involved in, and haplotypes predictive of, bone density, immune reactions, skin disease, and risk for various cancers [27], [28] and [29] with the FokI shown to be functional involved in the immune system [28]. The objective of the present study was to evaluate whether VDR polymorphisms are associated with the risk of idiopathic spontaneous PTB.

 

Materials and methods

The study was approved by the Institutional Ethical Board (Helsinki Committee) and the special Committee of the Israeli Ministry of Health for genetic studies.

 

Study design

This study was a 12-month prospective, case-control study. Subject enrolment and data collection were performed via the admission service of the Division for Maternal and Fetal Medicine in a large tertiary university-affiliated, Israeli obstetric department. Demographic data including maternal characteristics, past reproductive history, and information about previous complications during pregnancy, delivery, and the neonatal period were collected. Maternal and umbilical cord (fetal) blood samples from 33 Jewish and Arab (both ethnicities are considered Caucasian) mothers who had idiopathic spontaneous preterm birth (24–35 weeks gestation, but excluding multi-fetal pregnancy, fetal anomalies, polyhydramnios, prolonged PPROM, vaginal bleeding, clinical chorioamnionitis, uterine malformations, cervical cerclage, and birth related to coital activity) were compared to samples from 98 Jewish and Arab maternal and cord blood samples of term, uncomplicated births.

 

Mothers of full-term births were recruited sequentially within the same time-frame as those with a preterm birth: all those who signed informed consent were included up to a total of 98 women within 13 months from the start of the study. All mothers had benefited from unbiased medical care provided by the National Health Plan and had had at least four antenatal care visits. Pregnancies were dated by the date of the last menstrual period and/or a first trimester ultrasound.

 

PCR analysis

Maternal and umbilical cord blood samples (in 0.11 mol/l sodium tricitrate) were paired. DNA was extracted from peripheral blood cells using a high salt precipitation method and PCR amplification identified functional polymorphisms in the 5′ regulatory region of VDR gene by standard methods and subsequently stored at −4 °C for batched analysis [30]. Genotypes were repeated for (three) ambiguous cases.

 

Maternal and umbilical blood samples were coded and genotyped by standard methods with appropriate restriction endonucleases to identify SNPs: FokI, ApaI, TaqI, and BsmI ( Table 1) ( Fig. 1). For each polymorphism, five samples were sequenced for comparison. All results showed 100% matching between sequence and RLFP.

 

The laboratory staff was blinded as to the clinical status of the samples. The gel banding patterns were assessed independently by two of us (SG-G and GA).

 

Statistics

Demographic data, reproductive history, and information on complications during pregnancy, delivery, and the neonatal period are presented as descriptive statistics. Comparisons were performed with one-way ANOVA for continuous variables with post-hoc tests (Dunnett and Scheffe) correction for unequal variances. The χ2-test was used for categorical variables.

 

PHASE 2.0.2 was used to estimate allele frequencies and construct haplotype pairs for each subject. PHASE implements a Bayesian statistical method for reconstructing haplotypes [31]. All other analyses were performed with SAS 9.2. Hardy–Weinberg equilibrium was tested with χ2 goodness-of-fit for each SNP. The common allele frequencies were estimated with 95% confidence interval. Student's t-test, Fisher's exact test, and Pearson chi-square were used to test the association between the group pairs, and a p < 0.05 was considered significant. Odds ratios were calculated to assess the relationship between SNPs and haplotypes and the outcome. Odds ratio estimation under an additive genetic model was done by applying exact logistic regression.

 

Results

Maternal age, ethnicity (Jewish/Arab) and area of residence among all study groups were comparable. None of the mothers reported drug intake or abuse habits before or during the study period. All participants were medically insured under the National Health Plan and had similar antenatal care (including medical assistance and a pre-paid drug plan).

 

The PTB group differed from the mothers with healthy full-term babies because they had lower parity and a lower number of live births, a higher rate of cesarean deliveries, and a history of previous PTB. Cesarean deliveries in the PTB cohort were performed with established labor and none was due to a diagnosis of chorioamnionitis. Maternal and perinatal characteristics of both cohorts are presented in Table 2 and Table 3, respectively.

 

 

 

All four VDR polymorphisms Fig. 1 were in Hardy–Weinberg equilibrium for the cohort of mothers with healthy full-term babies. Maternal VDR wild type allele frequencies as described by perinatal outcomes are shown in Table 4. Maternal genotype frequencies (of the wild type allele) are presented in Table 5. Maternal FokI and TaqI VDR allele frequencies were significantly different in the PTB group (p = 0.01 and p = 0.0032, respectively): the FokI wild type (C allele) was increased in the mothers with PTB whereas the TaqI wild type (T allele) was decreased in the mothers with PTB. The other genotypes showed non-significant variations. There was no significant difference for either genotype in the umbilical cord samples. Multiple haplotype combinations analysis was not performed because of the limited number of samples in the study group. Thus, a logistic regression model based on the status of FokI and TaqI Table 1 as the functional alleles was devised to assess the additive risk models of the VDR SNPs by perinatal outcomes. In this type of additive analysis, the frequency of the allele has no impact. Using this approach, only maternal FokI variant was identified as associated with an increased risk for preterm birth with an odds ratio (OR) of 3.317 (95% CI: 1.143, 9.627; p = 0.0274, value for the maximum likelihood estimates). There were no comparably apparent increased ORs for the other three variants: ApaI variant OR of 0.413 (95% CI: 0.151, 1.135); TaqI variant OR of 0.232 (95% CI: 0.063, 0.862); and BSMI variant OR of 1.129 (95% CI: 0.434, 2.939).

 

 

Discussion

Spontaneous preterm labor and preterm births are still the leading causes of perinatal morbidity and mortality in the developed world [2]. Efforts to prevent preterm birth are hampered by inadequate understanding of the underlying pathophysiology, inadequate diagnostic tools, and generally, ineffective intervention because it is unclear whether all women should be targeted or whether there is a well-defined at-risk population [32]. It has been suggested that gene–environment interactions may better explain the risk of preterm birth, and if so, polymorphisms of critical genes might be used to ascribe risk [7]. We have therefore explored the possibility that VDR polymorphisms may serve as risk markers for PTB for two reasons: first, because the VDR system seems to be intimately involved at the feto-maternal interface, which is assumed to be the physical site of PTB [23], [33], [34] and [35]; and second, if nutritional status can ameliorate function, then intervention to improve outcome may be feasible [36] and [37]. Disappointingly, previous studies relating to the linkage between calcium metabolism and preeclampsia [38] and [39] did not lead to further investigations into the role of the VDR polymorphisms in clinical obstetrics [40].

 

In the present study, maternal FokI VDR allele frequency is associated with PTB, a relationship supported in an additive logistic model without dependence on the frequency of the allele in the population. The neonatal genotype (umbilical cord samples) showed no difference between PTB and control groups, implicating a maternal risk factor. These findings in part dovetail with those where maternal vitamin D status, modified by fetal FokI genotype was associated with infant birth size [41]. However, little can be implied as to the neonatal risk involved because of the very limited number of fetal observations. The other three polymorphisms in the VDR gene, BsmI, ApaI, and TaqI, at the 3′ end of the VDR gene, are probably non-functional, and in linkage disequilibrium with the functional FokI polymorphism. Furthermore, the FokI VDR variant has specifically been shown to correlate with immuno-modulatory [28] and anti-inflammatory features [42].

 

These results may speak to a “dual-hit” hypothesis wherein maternal VDR genotype results in early defective immuno-modulation at the placenta-decidua interface and, with pregnancy progression and consequent increasing vitamin D/calcium demands, PTB results. Genetic association studies, for detection of genetic variants contributing to a complex outcome such as PTB should indeed be interpreted with caution, but this speculation deserves more attention. In a study comparing normal placental tissue and placental tissue from premature births, there was a significant difference in VDR expression of mRNA levels (p = 0.041), with reduced VDR mRNA expression in those with PTB, indicating a possible dependence of the modulation of VDR expression on the proliferation and differentiation processes [43]. Potentially, down-regulation of VDR in placental cells may imply altered production of calcitriol, but this is unproven.

 

In the current study, while not under-powered, only 33 mothers with PTB were included: clearly a larger cohort would be preferable. In addition, while extraction was performed in a blinded manner, the only extraction failures were in the preterm neonates. This too may have ramifications of a qualitative nature as well as of a quantitative nature unique to this study. However, the sample size limitation of the present study is compensated for by some features of study design that may bolster the conclusions. Both groups are comprised of women who are similar in maternal age, socio-economic status, area of residence, prior reproductive history, antenatal care, and tertiary medical facility availability at birth, thus limiting the influence of known confounders described in other preterm births studies. All PTB cases were strictly selected, avoiding known associated factors for prematurity such as infections, thrombophilia, and maternal and fetal concomitant diseases. The homogeneity of ethnic backgrounds and area of residence, while confining our conclusions to an Israeli population, at the same time strengthens the genetic statistical analysis model. Nonetheless, we are cognizant of the approximately 26% loss of umbilical sample data with consequent exclusion from the haplotype analysis, making the neonatal cohort exceptionally small and inadequate for definitive conclusions.

 

Despite the aforementioned limitations of this study, this novel approach to PTB may assist in identifying and rationally counseling a select at-risk population by virtue of a complex “molecular signature” of the FokI VDR polymorphic genotype in the mother. Moreover, these findings have a practical aspect and might encourage a prospective study of the use of (nutritional) supplementation of vitamin D for pregnant women at risk for preterm birth. Similar recommendations have recently been made for women with recurrent pregnancy losses with low vitamin D levels [44]: however, the most provocative recent study [45] of a randomized community-based intervention suggests that vitamin D supplementation (2000 and 4000 IU/day) during pregnancy may reduce risk of infections, preterm labor, and preterm birth.

 

Condensation

Maternal FokI vitamin D receptor allele frequency was significantly different in preterm compared to full-term birth, suggesting FokI may be a maternal risk-trait for prematurity.

 

Disclosure

None of the authors has any financial conflicts of interest to report relevant to this study. No special funding was received for performance of this study.

 

Acknowledgments

The authors gratefully acknowledge the patience and expertise of Ms. Tali Bdolah-Abram, senior bio-statistician at the Hadassah Medical School of the Hebrew University for statistical assistance.