Correlation analysis of MTHFD1 gene polymorphisms and neural tube defects in Han population of Northern China

Neural tube defects (NTDs) is a common birth defects worldwide. The methylenetetrahydrofolate dehydrogenase1 (MTHFD1) gene has been proved to play an important role in folate metabolism, which was strongly associated with the increased NTDs risk. The study is aimed to explore the correlation of single nucleotide polymorphisms (SNPs) in MTHFD1 gene with NTDs susceptibility. A case-control study was conducted on children who included 152 NTDs patients and 169 healthy controls. Tag-SNPs were identied in HapMap database. Then, we investigated the association between NTDs and four selected tag-SNPs in MTHFD1 gene: rs1950902, rs2236225, rs2236224, rs11849530. We also performed a meta-analysis based on previous published studies to further evaluate the association.


Background
Neural tube defects (NTDs) are congenital malformations of the brain and spinal cord that usually occur in early pregnancy (21 28 days) [1], mainly including spina bi da, anencephaly and encephalocele. As reported, the prevalence is roughly 0.5 2 per 1000 births worldwide, which is higher in China with about 2.74 per 1000 births [2][3]. Related studies had revealed that pathogenesis of NTDs was believed to mainly involve genetic and environmental factors [4][5]. Furthermore, epidemiological studies testi ed that maternal folic acid supplement can reduce the incidence by 50 70%, suggesting that variants in genes involved in folate metabolism pathway may contribute to NTDs risk [6]. Methylenetetrahydrofolate dehydrogenase1 (MTHFD1) plays an important role in folate metabolism by catalyzing the conversion of tetrahydrofolate to the corresponding 10-formyl, 5,10-methenyl and 5,10-metheylene derivatives [7]. The MTHFD1 gene is located at chromosome 14q24 and its cDNA contains an open reading frame of 2805 bp that encodes a protein of 935 amino acids [8]. Related studies had indicated that MTHFD1 polymorphisms might be associated with increased NTDs risk, but the association remains controversial since several studies suggest no association.
So, the tag-SNPs of MTHFD1 gene were retrieved in our study, including rs1950902 (401A > G), rs2236225 (1958G > A), rs2236224 (2136 + 31G > A), rs11849530 (2458-2060A > G) and rs35020344 (41 + 239A > G) from the HapMap database based on related criteria. Then, we conducted a case-control study to assess the association between MTHFD1 polymorphisms and the risk of NTDs. Considering the reason for small sample size of an individual study, a further meta-analysis was performed so as to comprehensively evaluate the correlation.

DNA extraction and genotyping
Peripheral blood samples were collected from all participants in our study after obtaining their consent. The tubes used to collect peripheral blood contained EDTA as anticoagulation. Extraction of genomic DNA was carried out from blood samples using the Qiagen DNA Blood Mini Kit (Qiagen, UK). The collected DNA samples were stored at -80℃ before use. Five tag-SNPs of MTHFD1 gene were genotyped via the Sequenombased Mass ARRAY assay. Then, the data of genotype and allele distributions were incorporated and analyzed by the Filemaker Pro Database, which is a cross-platform relational database application from FileMaker Inc

Statistical analysis
All statistical analysis were performed by SPSS19.0 software. Statistical signi cance was accepted at P ≤ 0.05. The t-test and Chi-square test were used to estimate the difference in clinical characteristics, genotype and allele distributions between cases and controls. Hardy-Weinberg equilibrium was assessed by Chi-square in the control group in order to detect group representation. Then, meta-analysis of all relevant studies on the correlation between MTHFD1 G1958A polymorphism and NTDs risk was conducted by Stata 12.0 software. The pooled OR were calculated under ve genetic model to comprehensively assess the correlation. In metaanalysis, heterogeneity among studies was analyzed by Chi-square test-based Q-statistic. The pooled OR was calculated by xed effect model when P value was more than 0.1; otherwise, the random effects models was applied. Moreover, the online SHEsis software ((http://analysis2.bio-x.cn/myAnalysis.php) was used to analyzed the linkage disequilibrium (LD) and haplotype construction. D' and r 2 were calculated for LD analysis.

Study characteristic
In our research, 321 subjects were recruited, including 152 NTDs patients (87 males and 65 females) and 169 healthy controls (90 males and 79 females). As shown in Table 1, the average age in NTDs cases and healthy controls were 2.36 ± 1.21 and 2.41 ± 1.30, respectively, but the difference did not reveal any statistical signi cance (P = 0.722). Importantly, compared with healthy controls, NTDs cases appeared to have higher concentrations of homocysteine and SAH (P = 0.010 and P = 0.012). Moreover, the level of SAM in NTDs cases was lower than healthy controls (P = 0.021). In addition, we found a lower level of folate in NTDs cases compared with healthy controls, but the difference showed no statistical signi cance (P = 0.063).

Selection of tag-SNPs
Five tag-SNPs were identi ed from HapMap database according to above criteria. The LD plot of ve tag-SNPs was presented in Fig. 1. Because rs35020344 site did not conformed to Hardy-Weinberg equilibrium test, four tag-SNPs (rs1950902, rs2236225, rs2236224 and rs11849530) were selected by Haploview 4.0 to represent all SNPs in MTHFD1 gene. The SNPs of rs1950902 and rs2236225 were located in the exons, whereas the SNPs of rs2236224 and rs11849530 were situated in introns.

Association between haplotypes of MTHFD1 gene and the risk of NTDs
The haplotype was constructed to analyze the effect of the LD on NTDs in Han population of Northern China.
Our results indicated that there was a statistical signi cance in haplotype A-A, C-A and T-A between the two groups (P < 0.05). The OR with 95%CI of A-A and C-A haplotypes were 1.951(1.057-3.600) and 1.529(1.041-2.246), suggesting that they were harmful factors for NTDs. However, the OR with 95%CI of T-A haplotype was 0.682(0.497-0.93), indicating it was a protective factor for NTDs ( Table 2). Association between MTHFD1 polymorphisms and the risk of NTDs

Meta-analysis of the correlation between MTHFD1 1958G > A polymorphism and NTDs risk
Due to the limitation of sample size, we conducted a meta-analysis of association between MTHFD1 1958G > A polymorphism and the risk of NTDs based on previous studies. A total of seven case-control studies were identi ed according to the inclusion criterion [10][11][12][13][14][15][16].The characteristics of selected studies are summarized in Table 4. As shown in

Discussion
NTDs are a multifactorial disease and its precise mechanism remains unknown. Epidemiologic studies had con rmed the pathogenesis of NTDs was mainly involved in genetic and environmental factors. A number of studies had demonstrated that variants of several genes involved in the folate-dependent one-carbon metabolism had been shown to be associated with the risk of NTDs [17][18]. Besides, NTDs are associated with polymorphisms involved in planar cell polarity (PCP) signaling pathway [19][20][21]. Of these, the most important is the gene encoding MTHFD1, which plays an important role in one-carbon metabolism by providing folate cofactors for DNA synthesis and for cellular methylation reactions.
The aim of this study was to explore the association between MTHFD1 polymorphisms and the risk of NTDs. A total of 321 subjects including 152 NTDs cases and 169 healthy controls were enrolled in this study. Related studied showed that low S-adenosylmethionine (SAM), concentration in combination with increased Sadenosylhomocysteine (SAH) concentration had been supposed to reduce methylation capacity [22]. Moreover, many epidemiological studies further con rmed that mothers with NTDs offspring had lower folate and higher homocysteine than those with normal offspring [23]. In our study, comparing the clinical characteristics between case and control groups, there was a signi cant difference in homocysteine, SAM, SAH and SAM/SAH, which indicated that these factors might be responsible for the occurrence of NTDs.
Meanwhile, we selected four tag-SNPs (rs1950902, rs2236225, rs2236224 and rs11849530) to further evaluate the correlation between MTHFD1 polymorphisms and the risk of NTDs. Moreover, we also found that A-A (rs1950902-rs2236225) and T-A (rs2236224-rs11849530) haplotypes could increase the risk of NTDs, but C-A (rs2236224-rs11849530) haplotype might reduce the risk of NTDs.
To date, the MTHFD1 rs2236225 is a functional exonic SNP that has been extensively studied in several studies, but the results were inconsistent. Carroll et al reported no evidence concerning correlation between the rs2236225 polymorphism and the risk of NTDs [24]. But Van der Linden et al indicated that G1958A mutation was an important risk factor of NTDs [12].There could be several factors resulting in the con icting conclusions among individual study. The most likely reasons were the sample size of each study and selected different populations in each study. In our study, we found the rs2236225 polymorphism was signi cantly associated with increased risk of NTDs. In order to comprehensively evaluate the effect of MTHFD1 rs2236225 polymorphism on NTDs, we conducted a meta-analysis based on previous studies. Meta-analysis results showed a signi cant association between MTHFD1 rs2236225 polymorphism and the risk of NTDs, which further con rmed the reliability of our ndings. Besides, rs2236224 polymorphism is another common genetic variation in MTHFD1, we also found MTHFD1 rs2236224 mutation was associated with increased the risk of NTDs

Conclusions
In conclusion, our results revealed that the SNPs of rs2236225 and rs2236224 in MTHFD1 gene were associated with increased NTDs risk in Han population of Northern China. The meta-analysis in our study further supported the conclusion. Due to the limitation of this study, further study with a large sample and functional study should be conducted to get a deeper insight into the etiology of NTDs.

Declarations
The dataset and analyses are available from the corresponding author on reasonable request.
Authors' contributions F-YL, L-Y and S-JB analysed data and drafted the manuscript. Z-LS was responsible for clinical diagnosis.W-L and W-XT were responsible for experimental studies. Z-XF and Z-J were responsible for data collection. O-YS and C-CQ participated in the design and coordination of this study in addition to revising and critiquing the manuscript. All authors read and approved the nal manuscript.
Ethics approval and consent to participate This study was approved by the Tianjin Children' Hospital Ethics Committee. The guardian (parents) of the patient consented to both participation and publication of the case.

Consent for publication
Informed consent was obtained from the guardian (parents), who agreed to join this study, and using the medical information for scienti c research and publication.