Main

Preterm birth (PTB), defined as a birth before 37-wk gestation, is a major perinatal health problem. Approximately 10% of all births worldwide are preterm (1). There is wide variation by geography and socioeconomic status. Africa, Asia, and North America have the highest rates of PTB. PTB rate in United States was 11.4% in 2013 (2), significantly higher than other developed nations. PTB is directly responsible for an estimated one million neonatal deaths annually (3). It is also an important contributor to long-term morbidities such as cerebral palsy, developmental delay, visual and hearing impairments, and chronic lung disease (4,5).

The etiology of PTB is multifactorial and there is strong evidence for genetic susceptibility (6,7). PTBs tend to recur in mothers (8). Familial trends and racial disparities of prematurity also suggest that genetics influence this trait (9,10,11). A number of case–control and family-based studies have evaluated candidate genes chosen from some of the pathways associated with PTB (12,13,14,15,16,17).

In this study, we hypothesize that single nucleotide polymorphisms (SNPs) in four candidate genes (NR5A2, FSHR, FOXP3, and SERPINH1) contribute to genetic predisposition to PTB. We evaluated the association between PTB and SNPs in the four candidate genes.

The four candidate genes we studied act on fundamental mechanisms associated with the establishment and maintenance of pregnancy. NR5A2 gene encodes Nuclear Receptor Subfamily 5, Group A, Member 2, also known as liver receptor homolog-1 (LRH1). LRH1 has been shown to play an important role in establishing and sustaining pregnancy in animal models (18). Polymorphisms in FSHR encoding follicle-stimulating hormone receptor have been found to be associated with PTB in a Finnish as well as an African-American population (19,20). Dysregulation of FSHR may contribute to early uterine contractility. SERPINH1 encodes heat shock protein 47, which serves as a chaperone stabilizing the collagen triple helix. Polymorphisms in SERPINH1 increase risk of preterm premature rupture of membranes (PPROM) in African Americans (21). FOXP3 encodes a member of the FOX protein family that appears to function as a master regulator in the development and function of regulatory T cells an important pathway in modulating the immune interactions between mother and fetus. While many studies have focused on the fetus as the “risk case” in studies of PTB, in this study we used the mother as the risk case and a family-based approach using DNA from the mother and her parents to improve the chance for identifying risk alleles acting through the mother.

Results

We analyzed the genotype data of 728 maternal triads from United States, Argentina, and Denmark. Maternal triad refers to mother and maternal grandparents of a preterm infant. A description of the study population is provided in Table 1 . Data on ethnicity were only available for US triads (Caucasian 97%, African American 2%, and other 1%). A total of 24 SNPs in four candidate genes (NR5A2, FSHR, FOXP3, and SERPINH1) were studied ( Table 2 ).

Table 1 Demographic characteristics of study population
Full size table
Table 2 List of SNPs
Full size table

All SNPs tested had a minor allele frequency greater than 20%. Genetic predisposition for PTB was associated (P < 0.05) in multiple SNPs in our study as shown in Table 3 . Analysis of all maternal triads found a significant association of PTB with rs2737667 in NR5A2 (P = 0.02). Analysis of the US triads found a significant association of PTB with rs2737667 (P = 0.01) and rs603647 in SERPINH1 (P = 0.02). There were no significant associations in the maternal triads from Argentina or Denmark (see Supplementary Table S1 online).

Table 3 Significant SNPs
Full size table

When stratifying by gestational age, three SNPs in NR5A2 (rs12131233, P = 0.007; rs2737667, P = 0.04; and rs2816949, P = 0.02) and rs6165 in FSHR (P = 0.01) had P values <0.05 in the <32-wk gestational age group. There were no significant findings in the 32–36-wk group (see Supplementary Table S2 online).

When cases of PPROM were excluded rs2737667 in NR5A2 showed the strongest association with PTB (P = 0.0002) for US and Argentina maternal triads. This association was the only one to remain significant after correction for multiple testing (Bonferroni-corrected P value <0.0006). In the PPROM group, rs6165 in FSHR had a P value of 0.02 (Supplementary Table S3 online). Results are summarized in Table 3 .

Discussion

PTB is a common complex trait. The etiology of PTB depends upon the interplay between genetics and environmental factors such as nutrition, infection, stress, trauma, and drug use. In this study, we investigated the association of PTB and four candidate genes with different biological mechanisms. To the best of our knowledge, this is the first candidate gene study investigating the role of NR5A2 gene in the etiology of PTB.

A SNP in NR5A2, rs2737667 showed an association with PTB in all maternal triads, <32-wk gestation age group and in the spontaneous preterm with intact membranes group. Of these, association with spontaneous preterm with intact membranes reached formal level of significance (P = 0.0002). Two other SNPs in the same gene (rs12131233 and rs2816949) were associated with PTB in <32-wk group but did not reach formal level of significance.

The NR5A2 gene encodes an orphan nuclear receptor named LRH1. Orphan nuclear receptors regulate transcription independent of known ligands. LRH1 is expressed in high levels in the ovary with highest expression in granulosa cells and corpus lutea. It is also expressed in endometrium, liver, intestine, and pancreas. LRH1 plays a vital role in bile acid synthesis, cholesterol metabolism, and steroidogenesis (22). It also has a role in the regulation of progesterone synthesis (22). LRH1 mediates progesterone production in the ovary and its actions in the uterus. Lack of LRH1 activity deregulates the expression of genes involved in progesterone synthesis and metabolism. In one study, mice lacking NR5A2 in granulosa cells had extremely low progesterone levels (23). LRH1 in the uterus has shown to be essential for implantation, decidualization, and placentation in mouse models. Zhang et al. (18) concluded that LRH 1 is necessary for maintenance of the corpus luteum, for promotion of decidualization, and for formation of the placenta, therefore playing multiple roles in establishing and sustaining pregnancy. Progesterone is critical for maintaining pregnancy and decline of progesterone action is implicated in the onset of parturition. Prenatal administration of progesterone has significantly decreased the PTB in women considered to be at increased risk (24,25). We can speculate that women with polymorphisms in NR5A2 gene and resulting functional variations of LRH1 are at risk of PTB due to low progesterone levels and lack of progesterone-dependent actions on the uterus.

SNP rs2737667 is an intronic variant in NR5A2 gene located 15 bp from a noncoding regulatory element. There are no reports of clinical significance for this variant in the current literature. We can hypothesize that this intronic variant may affect the function of noncoding regulatory element and affect expression of LRH1. Changes in LRH1 expression may result in variations of cholesterol metabolism, steroidogenesis, and progesterone synthesis contributing to PTB.

One SNP in FSHR gene (rs6165) showed a P value <0.05 in the gestation age <32-wk group and PPROM group, but did not reach formal levels of significance once corrected for multiple comparisons. Notably our results did not replicate the previously described associations of FSHR polymorphisms and PTB (19,20). In addition, SNPs in SERPINH1 did not show an association with PPROM cases as shown previously (21). Previously described associations for both of these genes were in Finnish and African-American cohorts. It is possible that the causative alleles have different frequencies in different populations, which explains why we could not replicate the previous result in our study cohort. Genetic polymorphisms vary greatly between different populations, so as the etiologies of PTB. This explains why the same associations are not found across all populations.

Our study has several limitations. We did not study PTB in relation to different ethnicities as the data on ethnicity were only available for US triads. We had to exclude Denmark maternal triad from PPROM analysis due to the unavailability of PPROM status.

Notable strengths of the study include large sample size across three different regions of the world. We studied the true spontaneous PTB excluding indicated PTB and multiple gestations, leaving a clean cohort of PTBs. We further divided the cohort into spontaneous PTB with intact membranes and PPROM in our sub group analysis, as the mechanisms of PTB can be different in these two groups.

In conclusion, this study suggests a potential association between intronic SNPs in the NR5A2 gene and PTB. NR5A2 gene encodes for the LRH1 protein, which plays a critical role in regulation of cholesterol metabolism and steroidogenesis. It also has a role in the regulation of progesterone synthesis. These findings suggest that NR5A2 may be important in the pathophysiology of PTB and exploring noncoding regulators of NR5A2 is warranted.

Methods

Study Population

The study population consisted of 728 maternal triads (662 pedigrees and 2,049 individuals) from United States (four sites: University of Iowa Hospitals and Clinics in Iowa City, IA; Magee-Women’s Hospital in Pittsburgh, PA; University of Rochester Medical Center in Rochester, NY; and Wake Forest University in Wake Forest, NC), Argentina (two sites: Instituto de Maternidad y Ginecología Nuestra Señora de las Mercedes in Tucumán and Hospital Provincial de Rosario in Rosario), and Denmark (Island of Funen and the Danish National Birth Cohort) ( Table 1 ). Inclusion criteria were spontaneous PTB before 37 wk of gestation. Gestational age was based on best obstetrical estimate (last menstrual period or ultrasound examination). Multiple gestation and major fetal chromosomal or structural anomalies were excluded.

All individuals provided signed informed consent for study enrollment in accordance with the protocols approved by research ethics committees in the US (the University of Iowa Institutional Review Board (IRB), University of Pittsburgh IRB, University of Rochester Research Subjects Review Board, Wake Forest University Health Sciences IRB), Argentina (the Research Ethics Committee of Centro de Educación Médica e Investigaciones Clínicas), and Denmark (the Scientific-Ethical Committee of the Southern Danish Region and the Biomedical Research Ethics Committee of the Capital City Region of Denmark). DNA samples were obtained from the DNA repository at the University of Iowa. Patients were recruited between 1999 and 2013. Neonates and families in this DNA repository were recruited as part of an initiative to create a research bio bank of biologic material for subsequent use in investigations into the genetic contributions of PTB and neonatal diseases. DNA was originally extracted from venous blood, buccal swabs or saliva from the mothers and maternal grandparents. Demographic information was collected through medical chart abstraction.

SNP Genotyping

A total of 24 SNPs in four candidate genes (NR5A2, FSHR, FOXP3, SERPINH1) were studied ( Table 2 ). SNPs for FSHR and SERPINH1 were chosen based on previously reported associations with PTB. SNPs for FOXP3 and NR5A2 were selected using HapMap database to get 80% coverage of the gene with a minor allele frequency ≥20% for each SNP. SNPs were genotyped using Taqman probes (Applied Biosystems, Foster City, CA) and the Dynamic Array Integrated Fluidic Circuits (Fluidigm, San Francisco, CA). All SNP genotyping assays were available and ordered using the Assay-on-Demand service from Applied Biosystems. These genotyping assays included primers to amplify the region containing the SNP of interest and two TaqMan Minor Groove Binder probes that are specific to the polymorphic variant alleles at the site labeled with different fluorescent reporter dyes, FAM, and VIC. All reactions were performed using standard conditions supplied by Fluidigm. Following thermocycling, fluorescence levels of the FAM and VIC dyes were measured using the EP1 Reader and genotypes were scored using the Fluidigm Genotyping Analysis software.

Genotyping was completed on 728 maternal triads from Denmark (180 triads), Argentina (372 triads), and the United States (176 triads). Genotyping efficiency was >96% for all markers. Greater than 98% of samples tested provided usable genotyping information. Genotypes were entered into a laboratory database (Progeny, South Bend, IN) to generate datasets for analysis.

Statistical Analysis

Genotyping data were analyzed by transmission disequilibrium test using the Family Based Association Test software (FBAT, Cambridge, MA) (26,27) to identify nonrandom allele transmission from parents to offspring and P values were obtained from the maternal effect analysis. We were able to test for maternally mediated genetic effects using the maternal triad as the analysis unit. We analyzed the data for all maternal triads together as well as for each of the populations individually (US, Argentina, and Denmark). We then performed subgroup analysis for gestational age categories (<32 and 32–36 wk) and PPROM status (preterm labor with PPROM and without PPROM). The earliest gestational age was used to determine the gestation age for any mothers who experienced multiple PTB s. Mothers who experienced multiple PTBs and had discrepant PPROM status were excluded from analysis when conducting PPROM analyses. PPROM analysis was only completed for the US and Argentina maternal triads as PPROM data were not available for Denmark triads. To correct for the multiple tests performed in this study, a conservative Bonferroni correction would place significance at P < 0.0006 using a standard α of 0.05. However, given the exploratory nature of this initial study, less stringent values are also of interest.

Statement of Financial Support

The authors have no financial relationships relevant to this article to disclose.

Disclosure

The authors have no conflicts of interest to disclose.