Introduction

Preterm birth (PTB) is a common pregnancy-related disease defined as delivery before 37 weeks of gestation and is one of the major determinants of neonatal and maternal morbidity (Smith 2012). A previous study has reported that approximately 28% of all neonatal deaths are caused by PTB (Blencowe et al. 2012). In most developed countries, 5–7% of all births were reported as premature, and this figure appears to be increasing (Beck et al. 2010). The proportion of PTBs in the Korean population was 7.2% in 2016, which is 1.5 times more than that in 2006 (http://kostat.go.kr/portal/korea/kor_nw/1/2/1/index.board?bmode=read&aSeq=362574). Various factors, such as maternal anthropometrics, health condition, age, prenatal care, and socioeconomic status, have been associated with PTB (Han et al. 2011). However, the exact etiology of PTB is not yet well understood (Han et al. 2011).

Several recent studies have reported that homocysteine is associated with the development of PTB (Tellapragada et al. 2016; Wang et al. 2015). Homocysteine is a sulfur amino acid derived from the demethylation of methionine (Hankey and Eikelboom 1999). It can induce oxidative stress, DNA strand breakage, and apoptosis (Mattson and Shea 2003). An increased level of homocysteine may also cause trophoblast apoptosis and reduce gonadotropin secretion, which are related with abnormal placentation and PTB (Engel et al. 2006). It is reported that enzymes of the folate metabolism pathway can regulate the blood concentration of homocysteine (Bailey and Gregory 1999; Chen et al. 2016; Stanislawska-Sachadyn et al. 2010; Wu et al. 2017). In the folate metabolism pathway, the transcobalamin II (TCN2) transports vitamin B12 from blood to tissue for folate metabolism (Stanislawska-Sachadyn et al. 2010). Methylenetetrahydrofolate reductase (MTHFR) serves a key role in folate metabolism as it catalyzes the regeneration of 5,10-methylenetetrahydrofolate into 5-methyltetrahydrofolate, which transfers the methyl group to the vitamin B12 (Bailey and Gregory 1999). Vitamin B12 is a cofactor for the methylation of homocysteine to methionine by catalysis of the methionine synthase (MTR) and the coenzyme methionine synthase reductase (MTRR) (Wu et al. 2017).

Several studies have previously analyzed the association between the gene polymorphisms of enzymes in folate metabolism and the concentration of homocysteine (Kurzwelly et al. 2010; Mfady et al. 2014; Mohammadpour-Gharehbagh et al. 2018; Oussalah et al. 2017; Wu et al. 2013). Mohammadpour-Gharehbagh et al. (2018) showed that the MTHFR rs1537514 G allele is associated with the increased expression of the MTHFR enzyme that might cause downregulation of the homocysteine concentration. Moreover, Wu et al. (2013) showed that the MTHFR rs4846049 AA genotype reduces MTHFR protein levels and elevates the level of homocysteine. Kurzwelly et al. (2010) reported that the MTR rs1805087 GG genotype caused a decrease in MTR enzyme activity with higher homocysteine levels. Mfady et al. (2014) found that the MTRR rs1801394 AA genotype was related to 1.2-fold higher homocysteine concentration. In addition, the TCN2 rs1801198 polymorphism was significantly associated with increased homocysteine in the European population (Oussalah et al. 2017).

Here, we conducted a case–control study to investigate the genetic association between PTB and folate-metabolizing gene polymorphisms, including MTHFR rs4846049 (2572 C>A), rs1537514 (4869 C>G), MTR rs1805087 (2756 A>G), MTRR rs1801394 (66 A>G), and TCN2 rs1801198 (776 C>G) in Korean women. In addition, we evaluated the combined interaction among metabolizing gene polymorphisms.

Materials and methods

Subjects

We analyzed a total of 254 women recruited from the Gynecology Department at Dankook University Hospital in Korea. Of these samples, 111 patients with PTB were selected based on their gestational ages being < 37 weeks. The control group consisted of 143 pregnant women with no history of PTB or spontaneous abortion and at a gestation period of at least 38 weeks. Neither the patients with PTB nor those in the control group had any systemic disease, such as hypertension, gestational diabetes, coronary heart disease, placental abruption, chronic nephritis, multiple gestation, and fetal anomalies. All clinical interviews were conducted by an obstetrician, and informed consent was obtained from all the participants of this study. The study protocol was approved by the Ethics Committee of the Dankook University Hospital (date of approval and the project identification code are respectively 16 November 2017 and DKUH 2016-12-003-005).

DNA extraction and genotyping

Genomic DNA was extracted from buccal cells or peripheral blood with a GeneAll Exgene Clinic SV mini kit (GeneAll, Seoul, Korea). We used PCR–RFLP for genotyping the MTHFR rs4846049 and rs1537514, MTR rs1805087, MTRR rs1801394, and TCN2 rs1801198. Primer sets for the MTR rs1805087, MTRR rs1801394, and TCN2 rs1801198 were designed according to previous studies (Cai et al. 2010; Hozyasz et al. 2012). To design primers for the MTHFR rs4846049 and rs1537514, Primer3Plus was used (http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi). The PCR was performed in a total volume of 20 µl, comprising 10 ng of genomic DNA, 10 pM of each primer, 0.2 mM dNTPs, 2.0 mM MgCl2, 10 × PCR buffer, and 1.0 U NV DNA polymerase (NAVI BioTech, Cheonan, Korea). The PCR amplification was conducted with a C1000 Touch thermal cycler (Bio-Rad, California, USA) under the following conditions: 95 °C for 5 min, followed by 35 cycles at 94 °C for 1 min, each annealing temperature for 1 min, and then 72 °C for 1 min, before a final extension at 72 °C for 10 min. Each of the PCR products was digested with 1.0 U StyI-HF (MTHFR rs4846049) and Alw26I (MTHFR rs1537514), HaeIII (MTR rs1805087), NdeI (MTRR rs1801394), and MvaI (TCN2 rs1801198) restriction enzymes (Enzynomics, Daejeon, Korea) for 6 h at 37 °C. Table 1 shows each of the primer sets and the polymorphic restriction enzyme sites.

Table 1 PCR primer sequences, annealing temperature and restriction fragments size for each polymorphism
Full size table

Data analyses

An independent t-test was performed to compare the characteristic data (i.e., age, height, weight, pre-gestation weight, systolic blood pressure, diastolic blood pressure, birth weight, and gestational ages) using SPSS 21 Statistics (IBM Korea, Korea). A chi-square test was performed to assess the Hardy–Weinberg equilibrium (HWE). In addition, the odds ratio (OR) with 95% confidence intervals (CI) was calculated using the genotype and allele frequencies of cases and controls. A p value of < 0.05 was considered statistically significant. Furthermore, we analyzed the gene–gene interaction among the metabolizing gene polymorphisms using the Generalized Multifactor Dimensionality Reduction (GMDR) analysis (http://www.ssg.uab.edu/gmdr) to construct all possible combinations of the five polymorphisms (Chen et al. 2011). The Bonferroni correction was applied to adjust for multiple comparison tests. Statistical analyses were performed using the web-based statistics tools SISA (http://www.quantitativeskills.com/sisa/) and SNPstats (http://bioinfo.iconcologia.net/SNPstats) and calculated the statistical power using G*Power 3.1 (Faul et al. 2007).

Results

We analyzed a total of 254 pregnant women. There were no meaningful differences between the mean values of the ages, heights, pre- and post-pregnancy weights, systolic blood pressures, and diastolic blood pressures of the patients with PTB and those in the control group (p > 0.05). However, there were significant differences in the birth weights and gestational ages between the patients with PTB and those in the control group (p < 0.05) (Table 2).

Table 2 Characteristics data of patients with PTB (preterm-birth) (n = 111) and control groups (n = 143)
Full size table

The genotyping data of MTHFR rs4846049, rs1537514, MTR rs1805087, MTRR rs1801394, and TCN2 rs1801198 polymorphisms for the 111 patients with PTB and the 143 patients in the control group are presented in Table 3. The genotype frequencies of the five SNPs had no deviation from the HWE in patients with PTB and the control group patients (Table 3). MTR rs1805087 GG genotype (p = 0.031), and TCN2 rs1801198 CG genotype (OR 0.53, 95% CI 0.288–0.980, p = 0.042). In the genetic models of each polymorphism, we found significant differences between the patients with PTB and those in the control group (Table 3). In particular, the TCN2 rs1801198 polymorphism was observed as a protective factor for PTB in the dominant model (OR 0.55, 95% CI 0.310–0.982, p = 0.042). And we observed a marginal trend toward significance in MTHFR rs4846049 AA genotype (OR 0.15, 95% CI 0.018–1.205, p = 0.041).

Table 3 Genotype and allele frequencies of MTHFR, MTR, MTRR, TCN2 gene polymorphisms in the patients with PTB (preterm-birth) and control group
Full size table

We constructed possible genotype combinations of the folate-metabolizing gene polymorphisms (Table 4). As a result, MTHFR rs4846049/rs1537514, MTHFR rs1537514/MTRR rs1801394, MTHFR rs1537514/TCN2 rs1801198, and MTR rs1805087/TCN2 rs1801198 models were selected by GMDR analysis (cross-validation consistency ≥ 7/10). We found that the combination of the MTHFR rs1537514 CC genotype and the MTRR rs1801394 GG genotype is a risk factor for PTB (OR 3.14, 95% CI 1.155–8.555, p = 0.021). Meanwhile, the protective interactions were identified in the combination of the MTHFR rs1537514 CC genotype and the TCN2 rs1801198 CG genotype (OR 0.52, 95% CI 0.269–0.996, p = 0.047) and the MTR rs1805087 AA genotype and the TCN2 rs1801198 CG genotype (OR 0.49, 95% CI 0.248–0.992, p = 0.046). The p values were not significant after these differences were adjusted by the Bonferroni correction (threshold; p < 0.005).

Table 4 Genotype Combination Analysis for two polymorphisms of MTHFR, MTR, MTRR, TCN2 genes in the patients with PTB (preterm-birth) and control group
Full size table

Discussion

In this study, we evaluated the effects of folate-metabolizing gene polymorphisms on PTB. The results of this study show that folate-metabolizing gene polymorphisms, including MTHFR rs4846049, rs1537514, MTR rs4805087, MTRR rs1801394, and TCN2 rs1801198, are significantly associated with the development of PTB in Korean women (p < 0.05) (Table 3).

The GG genotype frequency of MTR 2576 A>G (rs1805087) was significantly different between patients with PTB and those in the control group (p = 0.031) (Table 2). The MTR rs1805087 polymorphism converts aspartic acid to glycine at codon 919, and the SNP is possibly associated with lower MTR activity, followed by increased homocysteine level (Haghiri et al. 2016). Previous studies have indicated that the MTR rs1805087 polymorphism has a significant association with gestational disorders (Kim et al. 2013; Sata et al. 2012; Shi et al. 2017). Kim et al. (2013) and Sata et al. (2012) both showed that the MTR rs1805087 is significantly associated with recurrent pregnancy loss (Kim et al. 2013; Sata et al. 2012). Similarly, the meta-analysis study observed that the MTR rs1805087 polymorphism is related with thrombophilia (Shi et al. 2017). These studies support our results, revealing the association between MTR rs1805087 and PTB.

The A allele of MTRR 66 A>G (rs1801394) polymorphism in the exon 2 of MTRR is associated with decreased homocysteine levels (Jones et al. 2018). For the MTRR rs1801394, our results indicate no association with PTB (p > 0.05) (Table 2). However, the MTRR rs1801394 GA genotype was reported as a risk factor of PTB (Engel et al. 2006), and Song et al. (2018) suggested that the MTRR rs1801394 polymorphism is associated with increased plasma folate and homocysteine in pregnant women. The TCN2 gene is located in 22q12.2, including 9 exons and 8 introns, and encodes the transcobalamin II (Regec et al. 1995). The CC genotype of TCN2 776C>G (rs1801198) polymorphism is a factor for higher B12 levels and a lower concentration of homocysteine (Stanislawska-Sachadyn et al. 2010). However, in this study, the CG genotype protected against PTB in Korean women (OR 0.53, 95% CI 0.288–0.980, p = 0.042) (Table 2). Previous studies have indicated that the TCN2 rs1801198 polymorphism has no association with recurrent spontaneous abortion in Iranian women (Hashemi et al. 2018) and recurrent implantation failure in Korean women (Kim et al. 2014; Park et al. 2019). To verify these inconsistencies, a larger sample set and functional studies are required.

The results of the MTHFR 2572 C>A (rs4846049) polymorphism showed that the AA genotype was not statistically significant, yet marginal tendency (OR 0.15, 95% CI 0.018–1.205, p = 0.041) (Table 2). It is reported that haplotypes of rs4846049 polymorphism and other MTHFR SNPs showed an association with recurrent pregnancy loss in Korean women (Kim et al. 2017). The rs4846049 polymorphism in the 3′ UTR region of the MTHFR is reported to be associated with the MTHFR protein level through modifying miRNA binding (Wu et al. 2013). The AA genotype has led a reduced MTHFR protein level compared with that of the CC genotype (Salehi et al. 2018). A decreased level of MTHFR was found to be associated with the development of hyperhomocysteinemia that is related with placental abruption (Ferguson et al. 2001).

MTHFR 4869 C>G (rs1537514) causes MTHFR mRNA overexpression (Mohammadpour-Gharehbagh et al. 2018). Increased MTHFR activity in vivo is associated with downregulation of plasma homocysteine and a protective factor of PTB (Micle et al. 2012; Roy et al. 2008). Mohammadpour-Gharehbagh et al. (2018) showed that the MTHFR rs1537514 G allele and the rs4846049 C allele protect against the abnormal development of the placenta. However, the results of the present study indicate no association between the MTHFR rs1537514 and PTB (p > 0.05) (Table 2).

This study also demonstrates the gene–gene interaction among folate-metabolizing gene polymorphisms (Table 4). We obtained conflicting results regarding PTB in the combination analysis for the MTHFR rs1537514 CC genotype. The combination with the MTRR rs1801394 GG genotype is associated with the development of PTB, whereas the combination of the MTHFR rs1537514 CC genotype and the TCN2 rs1801198 CG genotype has a protective effect against PTB. One potential explanation for this inconsistency is the various etiologies of PTB. Previous studies have demonstrated that there are several causes associated with PTB, including infection, inflammation, uteroplacental ischemia, hemorrhage, uterine overdistension, stress, and other immunologically mediated processes (Goldenberg et al. 2008). Furthermore, various gene polymorphisms are associated with PTB and interact differently with each other (Crider et al. 2005). According to previous studies, the result of our combination analysis of folate metabolic gene polymorphisms supports this explanation (Table 4).

The present study has its limitations. Firstly, the sample size is relatively small (Hattersley and McCarthy 2005). However, the statistical power of the sample size in this study is over 90%, which is higher than the standard of 80% (Hong and Park 2012). Secondly, as PTB is affected by the maternal and fetal genotype, a study that investigates the genetic association of both genotypes is needed (Wilcox et al. 2008). Unfortunately, we analyzed only the genotype data from mothers. In addition, many additional factors have been associated with PTB (Hong and Park 2012), but we did not consider gene-environment interactions, such as maternal smoking or alcohol consumption, low maternal body mass index, and advanced maternal age, which may have a significant effect on the association between folate metabolism and preterm birth (Muglia and Katz 2010).

Despite these limitations, this attempt is the first study to analyze the association between folate metabolic gene polymorphisms and PTB in Korean women. Moreover, we identified the gene–gene interaction in folate metabolic gene polymorphisms for PTB. Thus, our case–control study may provide pathological evidence for further studies.

In conclusion, our results suggest that folate metabolism gene polymorphisms, including MTHFR rs4846049, rs1537514, MTR rs1805087, MTRR rs1801394, and TCN2 rs1801198, have a significant association with the pathogenesis of PTB. In addition, we identified that these polymorphisms have gene–gene interaction in PTB. However, as PTB is multifactorial, a larger sample set and further functional studies are required.