DNA Methylation Status Of Developmentally Important Imprinted Genes In Human Pregnancy Loss After ART And Spontaneous Conception

Assisted reproductive technologies including in vitro fertilization （IVF）, intracytoplasmic sperm injection （ICSI）, in vitro maturation （IVM）, embryos freezing, and preimplantation genetics diagnosis （PGD） et al.. The offspring born after ART is becoming an important part of population, thus concerns about the safety of ART have been increasing. Recent studies suggest that in vitro fertilization （IVF）/intracytoplasmic sperm injection may be associated with increased incidence of Angelman syndrome （AS） and Beckwith-Wiedemann syndrome （BWS）, and Imprinting defect is one of the most important factors. The majority of ART-conceived children with BWS and AS demonstrated imprinting defects, so Concerns about imprinting disorders linked to ART are therefore increasing. ART may affect epigenetic regulation, particularly the expression of imprinted genes, thus affecting the embryo implantation, placentation, organ formation and fetal growth. The prevalence of miscarriage has been estimated as15-20%of all clinically recognized pregnancies. To some extent, the occurrence of early miscarriage reduced the birth effects. Some studies suggests that genomic imprinting defects affecting the maintenance of early pregnancy. In the present study, we analyzed the DNA methylation patterns of four imprinted genes including H19, KvDMRl, PEG1, and SNRPN in human abortions and stillbirths, comparing ART with spontaneous conception, In addition, the association between DNA-methyltransferase （DNMT）-3A and DNMT-3B promoter genetic variant and the susceptibility to pregnancyloss was evaluated.Part I Quantitative methylation analysis of developmentally important imprinted genes in human early pregnancy loss after ART and spontaneous conception[Objective]Animal studies have demonstrated that genomic imprinting defects may affect the maintenance of early pregnancy. ART procedures may affect epigenetic regulation, particularly the expression of imprinted genes, thus affecting the embryo implantation, placentation, organ formation and fetal growth. In the present study, we analyzed the DNA methylation patterns of four imprinted genes including H19, KvDMR1, PEG1, and SNRPN in human chrionic villus from early pregnancy loss, comparing ART with spontaneous conception, primarily evaluating epigenetic risks linked to pregnancy loss.[Materials&Methods]（1） Chorionic villus samples （CVS） were collected from women who underwent abortion procedures in Department of Gynecology and Obstetrics in Nanfang hospital from May2008to July2011. The duration of pregnancy ranged from6to11weeks, and the maternal age ranged from18to45years old. These patients had no known anatomic or genetic abnormalities.（2） Four groups were tested:spontaneous abortion after ART （SN, n=62）, multi-fetal reduction after ART （FA, n=73）, spontaneous abortion of natural pregnancies （SN, n=73）, and induced abortion of natural pregnancies （IN, n=69）.（3） Genomic DNA was extracted after sample collection by proteinase K digestion. Bisulfite treatment of genomic DNA was performed with the EpiTect Bisulfite Kit （Qiagen）. The DNA methylation patterns of H19, LIT1, and SNRPN were analyzed using methylation speeific PCR （MSP） and pyrosequencing, which was performed using PyroMark Q96ID System （Qiagen） and pyrosequencing reactions were performed according to the manufacturer’s instructions. The degree of methylation at each CpG site was determined by Allele Quantification （AQ） softwase-All samples were analyzed in triplicate. Bisulfite sequencing polymerase chain reaction （BSP） was then perform to confirm the pyrosequencing results.（4） SPSS16.0statistical analysis program was used to analyze the percentage of methylation in chorionic villus from the four groups. Box plots were generated using the program’s default parameters. The receiver operating characteristic （ROC） curve method was used to analyze the potential association between imprinted genes’ methylation percentage and the incidence of early spontaneous abortion. The study protocol was approved by the Institutional Ethics Committee of Nanfang Hospital, and informed consent was obtained from all patients.[Results]（1） None of the277chorionic villus samples showed a complete loss of maternal or paternal alleles of H149, KvDMR1, PEG1, and SNPRN. The PCR products from methylated maternal chromosome and those from unmethylated paternal chromosome have been detected in all samples.（2） The gestation and maternal age between samples from ART and spontaneous conceptions were not significantly different, neither were the samples from spontaneous abortions and normal early pregnancies. （3） Bisulfite conversion efficiency values for all samples were90.1%-100%. The mean methylation values for H19, KvDMR1, PEG1, and SNPRN were49.6±9.3%（range,30.3-91.7%）,49.0±9.4%（range,33.9-86.4%）,55.5±9.7%（range,38.5-89.4%）,41.3±4.3%（range,32.6-61.2%）, respectively. Clear hypomethylation （0%） or hypermethylation （100%） was not detected. The percentage methylation of all the studied genes for the four groups showed significant differences （P<0.05）. The percentage methylation in spontaneous abortion groups （SA and SN groups） was significantly higher than in the non-spontaneous abortion （normal early pregnancy） groups （FA and IN groups）（P<0.05）.（4） Box plot analysis showed that in the135spontaneous abortions,37of540（6.9%） analyzed DMR methylation values （for the four studied genes） represented outliers. In contrast, the142normal early pregnancies exhibited none outliers. Furthermore, the imprinted KvDMR1gene （20outliers,14.8%） exhibited significantly more potentially abnormal methylation values in spontaneous abortions than the other studied genes. Six of135（4.4%） spontaneous abortions displayed hypermethylated DMRs in two to three genes. It is interesting to note that all the outliers represented hypermethylated and no hypomethylated DMRs.（5） However, ROC curve analysis could not predict an increased risk of early spontaneous abortion related to the methylation percentage of the four genes in chorionic villus. In addition, The mean percentage methylation did not differ significantly between samples from ART group and from non-ART group （P>0.05）, neither were samples from in vitro fertilization and intracytoplasmic sperm injection （P>0.05）. There’s no evidence for an increased rate of potentially abnormal methylation values in the ART group.[Conclusions]（1） As some studies have suggested, imprinting errors of imprinted genes may contribute to early spontaneous abortion. The higher the methylation percentage, the greater the chance of early miscarriage.（2） ART procedures or different semination methods might not affect the methylation patterns of imprinted genes, which were potentially related to the early pregnancy loss.（3） The further work is to enlarge the sample size and gene quantity, confirming the hypothesis above.Part Ⅱ Quantitative methylation analysis of developmentally important imprinted genes in human stillbirths from ART and spontaneous conceptions[Objective]The development of fetal may end at any time during the whole pregnancy satge. Although the genetic diagnosis methods were more and more excellent, part of the pregnancy loss in second trimester were still with unknown aetiology. Recent studies showed that among many reasons for such miscarriages, gene imprinting defects are some of the possible causes. Animal and human studies suggested in vitro culture of embryos can cause methylation defects in individual genes, which might affect subsequent embryonic development and contribute to pregnancy loss. To evaluate the correlation between methylation of imprinted genes and pregnancy loss in second trimester, the methylation status of H19, KvDMR1, PEG1, and SNPRN were assessed using pyrosequencing and Bisulfite sequencing polymerase chain reaction （BSP）.[Materials&Methods]（1） Muscle samples （MS） were obtained from aborted fetuses and stillbirths in the Department of Gynecology and Obstetrics in Nanfang Hospital from August2010 to July2011. Gestational age of each fetus was determined by last menstrual period. The sex was identified by examination of the external genitalia. The duration of pregnancy ranged from18to26weeks, and the maternal age ranged from18to38years old. These patients had no known anatomic or genetic abnormalities.（2） Three groups were tested:stillbirths from ART pregnancy （n=13）, stillbirths from natural pregnancy （n=17）, and induced aborted fetuses from natural pregnancy （n=13）.（3） Muscle smples were dissected within2hours after abortion and genomic DNA was extracted after sample collection by proteinase K digestion. Bisulfite treatment of genomic DNA was performed with the EpiTect Bisulfite Kit （Qiagen）. The DNA methylation patterns of H19, KvDMR1, PEG1and SNRPN were analyzed using methylation specific PCR （MSP） and pyrosequencing. All samples were analyzed in triplicate. Bisulfite sequencing polymerase chain reaction （BSP） was then perform to confirm the pyrosequencing results.（4） SPSS16.0statistical analysis program was used to analyze the percentage of methylation in chorionic villus from the four groups. Box plots were generated using the program’s default parameters. The receiver operating characteristic （ROC） curve method was used to analyze the potential association between imprinted genes’ methylation percentage and the incidence of early spontaneous abortion. The study protocol was approved by the Institutional Ethics Committee of Nanfang Hospital, and informed consent was obtained from all patients.[Results]（1） None of the43muscle samples showed a complete loss of maternal or paternal alleles of H19, KvDMR1, PEG1, and SNPRN. The PCR products from methylated maternal chromosome and those from unmethylated paternal chromosome have been detected in all samples. （2） The gestation and maternal age between samples from ART and spontaneous conceptions were not significantly different, neither were the samples from spontaneous abortions and normal early pregnancies.（3） Bisulfite conversion efficiency values for all samples were92.6%-100%. The mean methylation values for H19, KvDMRl, PEG1, and SNPRN were53.1±3.8%（range,46.2-60.9%）,48.9±8.7%（range,39.3-83.4%）,53.1±3.8%（range,46.2-60.9%）,42.0±3.9%（range,36.4-54.8%）, respectively. Clear hypomethylation （0%） or hypermethylation （100%） was not detected. The percentage methylation of KvDMR1in stillbirths from natural pregnancy was significantly higher than in the stillbirths from ART pregnancy and induced aborted fetuses from natural pregnancy （P<0.05）. For the other three genes, no significant differences were found in the methylation percentage among three groups （P>0.05）.（4） Box plot analysis showed that in the30stillbirths,5of120（4.2%） analyzed DMR methylation values （for the four studied genes） represented outliers. In contrast, the13normal pregnancies in second trimester exhibited none outliers. And the outlier rates in H19, KvDMR1, PEG1, and SNPRN were4.7%（2/43）,2.3%（1/43）,4.7%（2/43）, and0（0/43）, respectively. All the outliers represented hypermethylated and no hypomethylated DMRs.1of43（2.3%） stillbirths displayed hypermethylated DMRs in two genes.（5） In addition, ROC curve analysis showed a positive correlation between the methylation percentage of KvDMR1and PEG1in muscle samples and rates of spontaneous abortion in second trimenster:the higher the methylation percentage, the greater the chance of miscarriage （P<0.05）. The mean percentage methylation did not differ significantly between samples from ART group and from non-ART group （P>0.05）, neither were samples from in vitro fertilization and intracytoplasmic sperm injection （P>0.05）. There’s no evidence for an increased rate of potentially abnormal methylation values in the ART group.[Conclusions]（1） Imprinting errors of imprinted genes may contribute to spontaneous abortion in second trimester. Especially for KvDMR1gene, the higher the methylation percentage, the greater the chance of miscarriage.（2） ART procedures or different semination methods might not affect the methylation patterns of imprinted genes, which were potentially related to the early pregnancy loss.（3） To confirm the hypothesis above, further work is also needed, such as enlarging the sample size, gene quantity, and different kind of tissue type.Part III The DNMT-3A and DNMT-3B promoter polyphisms and risk of human pregnancy loss[Objective]Methylation of gene promoters is one of the major regulatory mechanisms of gene expression and abnormal methylation patterns of imprinted genes have been demonstrated to be associated with pregnancy loss. The establishment of DNA methylation patterns requires de novo methylation that occurs predominantly during early development and gametogenesis. DNA-methyltransferase （DNMT）-3A and DNMT-3B are both essential for de novo methylation and play an important role in the development of embryogenesis and the generation of aberrant methylation in carcinogenesis. Here we hypothesized that genetic variants of the genes that are responsible for regulating genomic methylation are associated with increased risk of human pregnancy loss.[Materials&Methods]（1） Two types of tissue were collected from women who underwent abortion procedures in Department of Gynecology and Obstetrics in Nanfang hospital from May2008to July2011:（a） chorionic villus samples （CVS） and muscle samples （MS） from spontaneous abortions conceived by assisted reproductive technology （ART） and spontaneous way （study group）, n=152;（b） CVS and MS from normal early pregnancy and second trimester （control group）, n=155. The duration of pregnancy ranged from6to26weeks, and the maternal age ranged from18-45years old.（2） We selected one of the single nucleotide polymorphisms （SNPs）-448A>G in the DNMT-3A promoter region and one of the SNP-149C/T in the DNMT-3B promoter region, then genotyped subjects for the DNMT-3A and3B promoter polymorphism to determine the association between this genetic variant and risk of pregnancy loss. The distribution of-448A>G and-149C/T polymorphisms was detected in all samples.（3） Genomic DNA was extracted after sample collection by proteinase K digestion. The transition of A>G of DNMT-3A SNP creates a TaaⅠ restriction site, and for DNMT-3B the variant T allele had an AvrⅡl restriction site. PCR-restriction fragment length polymorphism （PCR-RFLP） was used to detect this A-G transition in the promoter of DNMT-3A at-448A>G （GenBank accession No. NT₀22184.14:g.4381840） and the C-T transition in the promoter of DNMT-3B at-149C/T （GenBank accession No. AL035071,46359C/T）. DNA sequencing analysis was performed to confirm the PCR-RFLP analysis. The study protocol was approved by the Institutional Ethics Committee of Nanfang Hospital, and informed consent was obtained from all patients.[Results]（1） The DNMT-3A genotypes AA, AG, and GG were detected in the spontaneous abortions and the controls. The genotyping by PCR-RFLP analysis was completely confirmed by DNA sequencing analysis. The allele frequency of-448A among pregnancy loss group and control group was34.2%（52/152） versus16.5%（25.5/155）, respectively. Overall, we found that, compared with GG homozygotes, the DNMT-3A-448AA homozygotes had an about16fold increased risk of pregnancy loss [OR=16.130,95%confidence interval （CI）,3.665-70.984] and AG homozygotes an OR of2.027（95%CI,1.247-3.293）.（2） When the analyses of DNMT-3A were stratified by the maternal age and gestation, the AA and AG genotypes were associated with an increased risk of pregnancy loss （for AA, OR16.897,95%CI=3.810-74.925; for AG, OR2.293,95%CI=1.371-3.834） at<12weeks, and the AA and AG genotypes were also associated with an increased risk of pregnancy loss （for AA, OR28.868,95%CI=3.734-223.156; for AG, OR1.900,95%CI=1.111-3.251） at<35years. However, the distribution of-448A>G in individuals derived from ART pregnancy was not statistically significantly compared with those derived from spontaneous pregnancy （P=0.661）.（3） The DNMT-3A genotypes AA, AG, and GG were then detected in the135chrionic villus samples from spontaneous abortions and142from controls （study population in Part I）. We found that the distribution of-448A>G in samples with outlier （n=29）was much higher than that in samples with non-outlier （25/29,86.2%vs.96/248,38.7%）（p=0.000）.（4） For DNMT-3B, we observed genotype frequencies of100%（TT） in study populations and control groups. The genotyping by PCR-RFLP analysis was also completely confirmed by DNA sequencing analysis.[Conclusions]（1） The DNMT-3A-448A>G polymorphism is a novel functional SNP and contributes to its genetic susceptibility to pregnancy loss, especially for early pregnancy loss （<12w） in younger women （<35years）. ART did not affect the distribution of of-448A>G in pregnancy loss and normal pregnancy.（2） At present, we cannot exclude the possibility that the observed TT genotype of DNMT-3B represent normal variation during fetal development.（3） Our study about the SNP of DNMT-3A and3B is restricted to the DNA level. Additional studies on the underlying molecular mechanism of this polymorphism and the expression of DNMTs are warranted.