| BackgroundThe pathophysiological state of the placenta is highly linked to the pathogenesis of pregnancy disorders.Pregnancy disorders,such as pre-eclampsia(PE)and placenta accreta spectrum(PAS),pose a serious threat to the mother and fetus’ lives,which are caused by developmental problems and functional impairment of placenta.Variations in placental DNA methylation may predispose to the onset of placental disease.DNA methylation is a significant type of epigenetic modification.Cell free DNA(cfDNA)derived from the placenta in the peripheral blood of pregnant women can be used as molecular biomarkers for non-invasive prenatal diagnosis to dynamically monitor placental development and function throughout pregnancy.However,There haven’t been any studies on the use of maternal peripheral blood cfDNA methylation to predict placental diseases like PE and PAS.ObjectiveFinally,a model for predicting preeclampsia and placental implantation was established by detecting placental differentially methylated gene fragments in peripheral blood cfDNA of pregnant women,combined with clinical information of pregnant women,and validated in clinical samples.Firstly,we used differentially methylated gene fragments to screen for possible molecular markers of placental disease after analyzing placental DNA methylation sequencing data in three maternal categories: normal control(NOR),preeclampsia(PE),and placenta accreta spectrum(PAS).Subsequently,primers for peripheral blood cfDNA amplification were designed and peripheral blood cfDNA methylation detection techniques were established.Finally,a model for predicting preeclampsia and placental implantation was established by detecting placental differentially methylated gene fragments in cfDNA and clinical information of case control study,and validated in prospective cohort study.Methods(1)By analyzing the placental DMR of PE and PAS,focusing on Located in the CGI and promoter regions,screen genes related to placental development and angiogenesis,use these genes to design primers,and build a technology platform for double standard Quantitive Methylation-Specific PCR(q MSP)technology and cfDNA methylation.We designed 8 sites as a panel and used LINE1 as an internal reference to detect the cfDNA methylation of 8 sites in the sample.(2)In this study,44 placental DNA samples were collected and extracted for RRBS sequencing.(14 in the control group,14 in the preeclampsia group,and 16 in the placenta accreta group)(3)By analyzing the placental DMR of PE and PAS,focusing on Located in the Cp G islands(CGIs)and promoter regions,screen genes related to placental development and angiogenesis,we design primers and build a double standard Quantitive Methylation-Specific PCR(q MSP)technology to detect cfDNA methylation in maternal plasma.We designed 8 genes as a panel and used LINE1 as an internal reference to detect the cfDNA methylation of in sample(4)We design a case-control study and a prospective cohort study.The disease prediction model is built using the data from the case-control research,and validate using the data from prospective cohort study.The case-control sample included 98 pregnant women(29 in the control group,33 in the PE group,and36 in the PAS group),and the methylation of cfDNA in peripheral blood was detected.A disease prediction model was constructed using logistic regression,combining peripheral blood cfDNA markers and maternal clinical indicators to construct.From the first trimester to the third trimester in Cohort study,we collected cohort study candidates’ peripheral blood samples for 7 time(marked as T1-T7).The cohort study included a total of 263 plasma samples from 41 pregnant women,and a total of 96 plasma samples from 15 people in the control group.Result(1)GO enrichment analysis was performed on the DMR of placental RRBS.The results showed that GO term angiogenesis appeared in the top 25 enriched biological processes in the three comparison groups,and GO term vasculogenesis was the top 20 most significantly enriched biological processes in the NOR vs PE group.Angiogenesis-related genes had significant DNA methylation changes in PE and PAS disease placentas.(2)The results of the case-control study showed that the methylation ratio of BMPR2(p<0.05),EPN2(p<0.01),STOX2(p<0.01)in the PE group was higher than that in the NOR group,and ITGAV(p<0.001),KLF5(p<0.01)decreased;the methylation ratio of EPN2(p<0.1),ITGAV(p<0.001),KLF5(p<0.1),and ADGRB3(p<0.001)in the PAS group was higher than that in the NOR group,the methylation ratio of BMPR2(p<0.01)decreased;the methylation ratio of BMPR2(p<0.001)and STOX2(p<0.05)in the PE group was higher than that in the PAS group,and the methylation ratio in BMP4(p<0.05),ITGAV(p<0.001),KLF5(p<0.001),ADGRB3(p<0.001)decreased.(3)The results of the cohort study showed that the methylation ratio of PAS was lower than that of NOR group,and the periods with statistical differences were as follows: EPN2 at T4(p<0.05)and T6(p<0.01),ITGAV at T1(p<0.05),T2(p<0.05),T4(p<0.01),T5(p<0.01)and T6(p<0.01),KLF5 in T2(p<0.05)and T3(p<0.05),ADGRB3 at T4(p<0.05)and T5(p<0.05).the methylation ratio of ITGAV and KLF5 at each time point were lower in the cohort samples compared with the NOR group,and the time points with statistical differences were as follows: ITGAV at T6(p<0.1),KLF5 at T5(p<0.1)0.05);in the cohort samples,compared with the NOR group,EPN2 was higher than the control before T4,and lower than the control after T4(T6 p<0.05);no significant difference was observed in ADGRB3.(4)Using the data of the case-control study,logistic regression was used to construct the disease prediction model and the ROC curve.The PE prediction model Model A included four methylation sites,and Model B included four methylation sites,as well as maternal age and pregnancy rate.Early BMI.The results showed that the AUC value of Model A in the case-control study was 0.904,and the AUC value of Model B was 0.931.The diagnostic performance of the PE prediction model was verified with the cohort sample data.The AUC value of Model B in the T1 period was 0.799,and the AUC value in the T6 period was0.818.The PAS prediction model Model A includes four methylation sites,and Model B includes four methylation sites as well as the age of pregnant women and the number of scar uteri in the past.The AUC value of Model A in the casecontrol study is 0.820,and the AUC value of Model B is 0.905.(5)The disease prediction model were built using logistic regression and the data from the case-control research.Four methylation sites were included in both Models A and B of the PE disease prediction model,and Model B bring in maternal age and early pregnancy BMI.In the case-control study,the results showed that the AUC value for Model A was 0.904 and for Model B was 0.931.The data from prospective cohort study.was used to validate the diagnostic efficacy of the PE prediction model,at T1 and T6 the AUC is 0.799 and 0.818 respectively.The PAS prediction model A included four methylation sites,model B included four methylation sites,maternal age and times of previous scarred uteruses.In case-control study,AUC of Model A and Model B is 0.820 and 0.905 respectively.Conclusion(1)According to the analysis of placental RRBS,placental DNA methylation modification in PE and PAS diseases has undergone significant changes.We noted significant changes in genes enriched in angiogenesis regulation.(2)The stability of cfDNA methylation detection technology is quit good.The average of coefficient of variation(CV)of each technical links was below 5%,which means this technology has high stability.(3)The results of peripheral blood cfDNA methylation detection showed that the methylation level of ITGAV and KLF5 decreased in the PE group of case-control study,and the methylation degree of EPN2 increased significantly.The methylation degree of ITGAV(T6,p<0.1)and KLF5(T4,p<0.05)decreased in the PE group in cohort study.(4)The PE disease prediction Model constructed with logistics has good prediction performance,and the PE prediction Model B has good prediction performance from early to late pregnancy(AUC ranges from 0.755 to 0.818),in which T1AUC=0.799 and T6 AUC=0.818.(5)This study is the first to report the use of maternal peripheral blood cfDNA thatl differentially methylated in placenta to predict the PE and PAS. |
PDF Full Text Request