| ObjectiveWe based on 16S rDNA second-generation high-throughput sequencing,to characterize the vaginal microbiota in early pregnancy,and the relationship with first-trimester pregnancy outcomes(Normal pregnancy NP,Spontaneous abortion SA,Tubal pregnancy TP),pregnancy locations(Intrauterine pregnancy IUP,tubal pregnancy TP),and TCM conditions(Deficiency condition D,Excessive condition E).MethodsWe collected the vaginal secretions of females who in early pregnancy(gestation age between 4-8 weeks),the first-trimester pregnancy outcomes and TCM conditions were determined by follow-up untiled then 14 completed weeks(98 days).The vaginal microbiota profiles were assessed by 16S rDNA second-generation high-throughput sequencing of the V3-V4 region.Firstly,we respectively described the vaginal microbiota at the level of phylum and genus.Secondly,Alpha diverity community characteristics including PD-whole-tree,Pielou evenness,Shannon-Wiener diversity index as well as Beta diversity metrics including PcoA and Anosim base on Unweighted Unifrac and Weighted Unifrac were compared in different first-trimester pregnancy outcomes and TCM conditions.Thirdly,LefSe were used to find the vaginal biomarkers in different first-trimester pregnancy outcomes and TCM conditions.Fourthly,CCA were used to analyze the relationship among first-trimester pregnancy outcomes,vaginal microbiota and serum hormone.Futhermore,heatmap showed the clustering of the relative abundance of genera,first-trimester pregnancy outcomes and TCM conditions.Based on above research,genera were selected to predict the target classes named TP and IUP.To classify the different samples,2 models were applied,one was 10-fold cross-validation error in CART,the other was ROC(bootstrap resampling 500 times).Results1.Cohort and specimen characteristics:Finally,100 participants were included the analysis.46 participants were NP(31 were NPD and 15 were NPE),17 were SA(10 were SAD and 7 were SAE),37 were TP(7 were TPD and 30 were TPE).All samples were sequenced to sufficient depth.2.Characteristics of vaginal microbiota on total cohort:The Rank-Abundance curve suggested significant difference on richness among samples,and most samples had dominant OTUs.At the level of phylum,the total cohort was mainly composed of Firmicutes,Actinobacteria and Bacteroidetes.At the level of genus,the total cohort was mainly composed of Lactobacillus,Gardnerella,Atopobium,and Prevotella.Rarefaction curves showed sequencing could basically reflect the diversity of all samples.3.Characteristics of vaginal microbiota among different first-trimester pregnancy outcomes and pregnancy locations:The proportions of Firmicutes(91.08%,91.64%)and Lactobacillus(89.21%,88.40%)were prominent dominant in NP and SA.The proportions of Firmicutes(75.11%)and Lactobacillus(67.75%)in TP were not prominent.There was no significant difference in Alpha diversity index and Beta diversity among NP,SA and TP(P>0.05).As for IUP and TP,in Alpha diversity index,shannon-wiener diversity index of TP =1.1(1.0-2.3),IUP=1.0(0.4-1.7),P=0.04.In Beta diversity metrics,weighted anosim:R=0.14,P<0.01.Based on LEfSe,vaginal biomarker of NP was Lactobacillacea.Vaginal biomarker of SA was Bacillales.Vaginal biomarker of TP was FusobacteriiaFusobacteriales-Leptotrichiaceae,Actinobacteria,BifidobacterialesBi fi dobacteriaceae.Based on CCA,vaginal microbiota in NP had closest relationship with serum hormone among 3 different frist-trimester pregnancy outcomes.Serum hormone had a great effect on vaginal microbiota in NP from the largest to the smallest were progestin(P),estradiol(E2)and human chorionic gonadtrophin(β-hCG).4.Characteristics of vaginal microbiota between deficiency condition and excessive condition:The proportions of Firmicutes(88.24%)and Lactobacillus(85.75%)in deficiency condition were more prominently than those in excessive condition(82.53%and 76.87%).In Alpha diversity index,the total OTU in excessive condition was 100.5(68.8-145.2),and in deficiency condition was 78.0(54.8-108.0),P=0.02.Pd-whole-tree in excessive condition=8.1(6.2-10.0),in deficiency condition=7.1(5.7-8.5),P=0.03.There was no significant difference in Beta diversity metrics,P>0.05.Based on LEfSe,vaginal biomarker of excessive condition were Actinobacteria-Corynebacteriales-Corynebacteriaceae,Campylobacteria-Campylobacterales-Campylobacteraceae,Staphvl ococcaceae,Bacillal es,Rhizobiaceae,Del tapro teobac teri a,Xanthomonadales-Xanthomonadaceae.5.Characteristics of vaginal microbiota among different first-trimester pregnancy outcomes combined with TCM conditions:At the level of phylum,the proportions of Firmicutes were:NPD:90.53%,NPE:92.22%:SAD:94.68%,SAE:87.31%;TPD:68.87%,TPE:76.56%.At the level of genus,the proportion of Lactobacill were:NPD:88.11%,NPE:91.49%;SAD:93.99%,SAE:80.41%;TPD:63.48%,TPE:68.74%.In Alpha diversity index,total OUT:SAD:69.5(51.8-94.5),SAE:120.0(110.0-151.5),P=0.01.Pd-whole-tree,SAD:6.9±1.6,SAE:10.0±3.2,P=:0.02.In Beta diversity metrics,between SAD and SAE,unweighted anosim:R=0.23,P=0.01;weighted anosim:R=0.18,P=0.047.Clustering of individual samples showed 2 groups depending on abundance of Lactobacillus OTUs.One was the Lactobacillus dominated microbiota,had lower diversity;the other was non-Lactobacill us dominated microbiota,had higher diversity.IUP clustered at Lactobacillus dominated microbiota,TP clustered at non-Lactobacillus dominated microbiota.6.Non-Lactobacillus dominated microbiota(NLDM)and Lactobacillus dominated microbiota(LDM):7 genera(relative abundance in TP>0.01)were selected to predict the target classes named IUP and TP.Through Curve Fitting,there were 6 genera with positive correlation with TP:Gardnerella,Prevotella,Atopobium,Sneathia,Megasphaera,Enaerococcus.Lactobacillus negative correlation with TP.We established ROC(bootstrap resampling 500 times)of TP based on the relative abundance of 7 genera one by one.The AUC(95%confidence interval)from the largest to the smallest were:Lactobacillus was 0.70(0.58,0.79),Gardnerella was 0.68(0.58,0.78),Enaerococcus was 0.66(0.53,0.76),Megasphaera was 0.65(0.48,0.78),Prevotellawas 0.64(0.48,0.78),Sneathia was 0.62(0.48,0.81),and Atopobium was 0.61(0.45,0.74).Lactobacillus had the best prediction effect on TP with the highest accuracy(0.73).The best threshold of Lactobacillus relative abundance was 0.88.10-fold cross-validation error in CART provided similar conclusion:if Lactobacillus relative abundance≥0.88,the classification is IUP;while if Lactobacillus relative abundance<0.88,the classification is TP.The AUC for CART was 0.78.Then,based on these classifications,an vaginal microbiota profile—NLDM or LDM,was assigned to each participants.There were no statistically significant among age,BMI,tobacco use,gravidity,previous ectopic pregnancy,previous surgery history,plvic infection,gestationanl age and TCM condition between NLDM and LDM.Although participants with LDM(43.1)showed higher early pregnancy loss rate than NLDM(17.9%),P=0.02;but there was no statistically significant on pregnancy loss(all trimesters).In Alpha diversity index,pd-whole-tree:NLDM=8.3(7.5-9.5),LDM=7.0(5.3-8.9),P<0.01;Pielou evenness:NLDM=0.56(0.39-0.69),LDM=0.22(0.05-0.26),f<0.01;Shannon-wiener diversity index:NLDM=2.6(1.8-3.0),LDM=1.0(0.2-1.2),P<0.01.In Beta diversity metrics,weighted Anosim:R=0.12(P=0.04).In the aspect of vaginal environment,there was no difference between NLDM and LDM in vaginal bleeding,H2O2,LEU,β-GD,AG(P>0.05).pH was lower in LDM(4.6(4.2,4.6))than NLDM(4.6(4.6,5.0)),P<0.01.Sialidase positive rate was higher in NLDM(35.7%)than LDM(6.9%),P<0.01.In the aspect of serum hormone,β-hcg:NLDM=1972.5(253.0-13969.5)mIU/mL,LDM=8968.0(1393.5-56514.0)mIU/mL,P=0.03;P:NLDM=8.6(4.3-18.0)ng/ml,LDM=17.0(7.2-25.3)ng/ml,P=0.01;E2:NLDM=110.2(35.2-385.8)pg/ml,LDM=418.6(144.8-797.3)pg/ml,P<0.01.Conclusion1.There were great differences in the vaginal microbiota among individuals in early pregnancy,but Firmicutes and Lactobacillus dominated most of them.2.Based on LEfSe,vaginal biomarker of NP was Lactobacillacea.Vaginal biomarker of SA was Bacillales.Vaginal biomarker of TP was FusobacteriiaFusobacteriales-Leptotrichiaceae,Actinobacteria,BifidobacterialesBifidobacteriaceae.3.There were no significant difference among NP,SA and TP in early pregnancy vaginal microbiota.4.Vaginal microbiota in NP had closest relationship with serum hormone among 3 different frist-trimester pregnancy outcomes.Serum hormone had a great effect on vaginal microbiota in NP from the largest to the smallest were P,E2 and β-hCG.5.In early pregnancy,the diversity of vaginal microbiota in TP was slightly higher than in IUP.6.Vaginal biomarker of excessive condition were Actinobacteria-Corynebacteriales-Corynebacteriaceae,Campylobacteria-Campylobacterales-Campylobacteraceae,Staphylococcaceae,Bacillales,Rhizobiaceae,Deltaproteobacteria,Xanthomonadales-Xanthomonadaceae.7.In early pregnancy,the richness of vaginal microbiota in excessive condition was significantly higher than that in deficiency condition,but the structure of the two conditions were similar.8.In early pregnancy,the differences of vaginal microbiota between different conditions were mainly reflected in SA.In SA,the richness of vaginal microbiota in excessive condition was significantly higher than that in deficiency condition.9.In early pregnancy,the relative abundance of Gardnerella,Prevotella,Atopobium Sneathia,Megasphaera and Enaerococcus positively correlated with TP.While the relative abundance of Lactobacillus was negatively correlated with TP.10.In early pregnancy,the relative abundance of Lactobacillus could predict the occur of TP.If Lactobacillus relative abundance≥0.88,called LDM,the classification is IUP;while if Lactobacillus relative abundance<0.88,called NLDM,the classification is TP.11.Vaginal bleeding did not affect the classification of NLDM and LDM.LDM has higher relative abundance of lactobacillus,lower pH of vaginal secretions and higher levels of β-hcg,P,E2.NLDM with higher diversity and highter sialidase positive rate.However,NLDM did not have more severe vaginal mucosal injury. |
PDF Full Text Request