Date Published: May 10, 2017
Publisher: Public Library of Science
Author(s): HyungChul Rah, Ki Wha Chung, Ki Han Ko, Eun Sun Kim, Jung Oh Kim, Jung Hyun Sakong, Ji Hyang Kim, Woo Sik Lee, Nam Keun Kim, Bernard Mari.
MicroRNAs (miRNAs) regulate gene expression during the peri-implantation period. The purpose of this study was to investigate whether genetic polymorphisms in the four miRNAs associated with fetal or placental development play roles in the development of idiopathic recurrent pregnancy loss (RPL) in Korean females.
A case-control study involving 225 controls and 387 women with at least two consecutively recurrent pregnancy losses between 1999 and 2012 was performed. The genotypes of the four miRNA polymorphisms, including miR-27a rs895819, miR-423 rs6505162, miR-449b rs10061133, and miR-605 rs2043556, were analyzed by the polymerase chain reaction-restriction fragment length polymorphism assay. Odds ratios and 95% confidence intervals were estimated using multivariate analyses after maternal age adjustments. The relationships between each of the four microRNA genotypes and each of the six clinical parameters of the RPL patients (plasma homocysteine and folate levels, natural killer cell number, platelet count, prothrombin time, and, activated partial thromboplastin time) were analyzed using multiple linear regression analyses.
Our results suggest that weak associations between decreased RPL risk and the genotypes of miR-27a (AG and AG+GG), combination genotype of miR-27a/miR-423 (AG/GC), and haplotypes of miR-27a/miR-423/miR-449b/miR-605 (G-C-A-G) and miR-27a/miR-449b/miR-605 (G-A-G), whereas weak associations between increased RPL risk and genotypes of miR-449b (GG and AG+GG), combination genotypes of miR-423/miR-449b (CC/GG and CA/AG), miR-449b/miR-605 (AG/AG), haplotypes of miR-27a/miR-423/miR-449b/miR-605 (A-C-G-A, A-A-A-G, and G-C-G-G), miR-27a/miR-423/miR-449b (A-C-G), miR-27a/miR-449b/miR-605 (A-A-G, A-G-A, and G-G-G), miR-423/miR-449b/miR-605 (C-G-G and A-A-G), and miR-423/miR-449b (C-G and A-A). The genotypes of miR-27a (AG and AG+GG) also showed significant contributions to the prediction of folate levels in RPL patients.
The study showed associations between miRNA polymorphisms (miR-27a rs895819 and miR-449b rs10061133) and RPL development, and between the miRNA polymorphism (miR-27a rs895819) and plasma folate levels.
Recurrent pregnancy loss (RPL) or recurrent spontaneous abortion has been defined as the occurrence of at least two consecutive pregnancy losses prior to the 20th week of gestation [1, 2]. RPL occurs in approximately 1% of all pregnancies; however, the etiology for more than half of the RPLs remains undetermined . Genetic variation has been suggested one of the contributing factors leading to RPL and a number of single nucleotide polymorphisms (SNPs) have been reported to be associated with RPL . MicroRNAs (miRNAs) are short (approximately 22 nt) noncoding RNA molecules regulating expression of target genes at the post-transcriptional level by translational repression or messenger RNA degradation . Several studies recently reported the associations between miRNA polymorphisms and RPL [6–9]. One study identified two SNPs in miR-125a altering the production of miR-125a which was subsequently associated with an elevated risk for RPL in the Han Chinese women . Another study reported an association between two pre-miRNA polymorphisms (miR-196a2 and miR-499) and the occurrence of RPL in Korean females , which was supported in Iranian women . The most recent study identified a polymorphism in the coding region of miR-423 contributing to an increase in the expression of mature miR-423 associated with RPL in the Han Chinese population . Several miRNAs that are considered important during pregnancy were chosen for this study because of their elevated expression (miR-27a), decreased expression in the endometrium and in trophoblasts (miR-423), lower expression during endometriosis (miR-449b), and involvement in pregnancy loss via the p53 network (miR-605) [10–13]. In this study, we determined the susceptibility to RPL associated with genetic variants of miRNAs associated with placental or fetal development.
The demographic characteristics and clinical profiles of RPL patients and control subjects are shown in Table 1. The two groups were matched for age and BMI. Platelet numbers were significantly higher in the patient group than in the control group. The genotype and allele frequencies of the four miRNA SNPs in females with RPL and controls are shown in Table 2. All genotypes in the study cases were in HWE. In Table 2, the miR-27a A>G polymorphism was significantly associated with a risk of RPL (AA vs. AG: AOR = 0.654; 95% CI = 0.456–0.937; AA vs. AG+GG: AOR = 0.682; 95% CI = 0.484–0.960); however, there was no association after adjustment for multiple tests using the FDR correction. When RPL patients were stratified according to the occurrence of consecutive recurrent pregnancy losses (RPL = 2 and ≥ 3 vs. all RPL patients with RPL ≥ 2), there was an association between the miR -27a A>G polymorphism and RPL risk in the RPL ≥ 3 subgroup alone (AA vs. AG: AOR = 0.611; 95% CI = 0.404–0.923; AA vs. AG+GG: AOR = 0.639; 95% CI = 0.432–0.945). However, the associations were not statistically significant after adjustment for multiple tests using the FDR correction. The miR-449b A>G polymorphism was significantly associated with RPL risk (AA vs. GG: AOR = 2.069; 95% CI = 1.033–.4.146; AA vs AG+GG: AOR = 1.406; 95% CI = 1.011–1.955). However, the association was not significant after adjustment for multiple tests using the FDR correction. Using combination analyses (Table 3), the AG/CC (AOR = 0.579; 95% CI = 0.366–0.917) combined genotype for miR-27a/miR-423 was associated with a lower RPL risk compared with reference genotypes when variant genotypes were located in the miR-27a loci. In addition, CC/GG (AOR = 2.888; 95% CI = 1.116–7.470), CA/AG (AOR = 1.925; 95% CI = 1.110–3.338) for the miR-423/miR-449b, and AG/AG (AOR = 1.804; 95% CI = 1.067–3.052) for the miR-449b/miR-605 were associated with an increased RPL risk compared with reference genotypes when variant genotypes were located in the loci of miR-423, miR-449b, and miR-605. These results were consistent with associations between a RPL risk and individual microRNA genotypes; however, each association was not significant after the FDR correction for multiple comparisons, suggesting a weak association. Haplotype-based analyses of the four microRNA polymorphisms for gene-gene interactions are shown in S1 Table (all possible allele combinations) and Table 4 (allele combinations suggesting associations with RPL). Interaction models suggested by the MDR were evaluated using haplotype-based analyses. Among the models of the four polymorphic loci, the G-C-A-G haplotype (OR = 0.525; 95% CI = 0.321–0.859) was associated with a decreased RPL risk whereas three haplotypes, A-C-G-A (OR = 1.870; 95% CI = 1.178–2.968), A-A-A-G (OR = 2.429; 95% = 1.153–5.114), and G-C-G-G (OR = 3.214; 95% CI = 1.441–7.172), were associated with an increased RPL risk. Among the models of the three polymorphic loci, one haplotype, G-A-G of miR-27a/miR-449b/miR-605 (OR = 0.625; 95% CI = 0.402–0.972), was associated with a reduced RPL risk, whereas haplotypes A-C-G of miR-27a/miR-423/miR-449b (OR = 1.498; 95% CI = 1.030–2.179); A-A-G, A-G-A, and G-G-G of miR-27a/miR-449b/miR-605 (OR = 1.526; 95% CI = 1.037–2.244; OR = 1.649; 95% CI = 1.086–2.504; OR = 3.089; 95% CI = 1.401–6.809), C-G-G, and A-A-G of miR-423/miR-449b/miR-605 (OR = 1.773; 95% CI = 1.110–2.833; OR = 2.253; 95% CI = 1.203–4.220) were associated with an increased RPL risk. Among the models of the two polymorphic loci, two haplotypes, C-G and A-A of miR-423/miR-449b (OR = 1.518; 95% CI = 1.134–2.031), were associated with a higher RPL risk. The G-G-G of miR-27a/miR-449b/miR-605 and C-G and A-A of miR-423/miR-449b haplotypes remained significant after adjustment for multiple tests using the FDR correction. Multiple linear regression analyses of clinical variables in Korean RPL patients according to the quintiles of clinical variables are shown in Table 5. The AG and AG+GG genotypes of the miR-27a polymorphism showed significant contributions to the prediction of folate levels in RPL patients, with regression coefficients of 1.069 and 0.788, respectively. Although the differences in plasma homocysteine, folate, PLT, PT, aPTT, and NK cell number in relation to the four microRNA genotypes and haplotypes were evaluated by one-way analysis of variance and independent two sample t-tests, no significant difference was found (S2 and S3 Tables).
The effects of miRNA polymorphisms on pregnancy loss have been reported in a limited number of studies [6–9]. These studies investigated the roles of miR-125, miR-196a2, miR-499 and miR-423 polymorphisms that were supported by functional evidence showing disruption of mature microRNA production and its downstream target gene [6–9, 19]. Because more evidence in support of the functional importance of miRNAs in pregnancy has been reported, we determined whether the four miRNA SNPs associated with placental or fetal development played a role in pregnancy loss [14, 20–27]. Our results showed an association between the miR-27a variant G allele and a lower RPL risk, and an association between the miR-449b variant G allele and a higher RPL risk. The effects of variant alleles of miR-27a and miR-449b were also suggested using combination and haplotype-based analyses. Multiple linear regression analyses of clinical variables in Korean RPL patients revealed statistically significant relationships between the miR-27a genotypes and plasma folate levels. Functional analyses indicated that the variant genotypes of miR-27a, AG, and GG might be responsible for the elevated miR-27a levels , and it has been speculated that the elevated miR-27a levels contribute to the increased folate concentration that is protective against RPL . However, the possible effects of mir-27a on RPL mediated by folate require further investigation to confirm this hypothesis. The miR-449b was chosen for the study because it is one of the four miRNAs that was downregulated in hatched blastocysts . However, the effect of rs10061133 A>G of miR-449b has rarely been reported, and there have been contradictory reports on its effects, depending on the cell type and gene expression pattern . In our study, we found that the GG and AG+GG genotypes of miR-449b were associated with an increased risk of RPL. We, therefore, hypothesize that the GG and AG+GG genotypes of miR-449b rs10061133 affect the risk of RPL risk by modulating the expression of mature miR-449b [24, 27].