Table 1.
Clinical characteristics.
Fig 1.
Characterization of plasma EVs (EVs).
A. Immunoblot showing enrichment of CD63 (~60 kDA) in EVs compared to plasma, supernatant (Sup), and placenta and enrichment of placental-specific placental alkaline phosphatase (PLAP; ~70 kDA) (top panel). Immunoblot showing absence of cell-specific marker Calnexin (~90 kDA) in EV and plasma samples and presence in placenta and HepG2 cell line (bottom left panel) and enrichment of flotillin, an EV-specific marker in EVs compared to plasma. B. Size distribution and intensity of EVs detected by dynamic light scattering (red, green, blue peaks represent replicates) with table below showing the average size and intensity. C. Representative transmission electron microscopy of EVs (+EVs) with a negative control (-EVs), scale bar = 200 nm. Black arrows demonstrate the EVs.
Fig 2.
Differentially abundant EV-associated miRNAs at three trimesters of gestation and delivery in normal and gestational diabetes mellitus (GDM) pregnancies.
Venn diagrams display the number of differentially abundant miRNAs in EVs isolated from normal pregnant (NORM) or gestational diabetes mellitus (GDM) samples compared to non-pregnant (NP) samples in (A) first trimester, (B) second trimester, (C) third trimester and (D) at delivery. Yellow represents Normal pregnancy and blue GDM. All represented genes had a log2 fold-change ≥ 1 or ≤ −1 and FDR < 0.05. The FDR was calculated in DESeq2 using the Benjamin-Hochberg correction [22].
Fig 3.
Distribution of EV derived miRNAs at three trimesters of gestation and at delivery in normal and gestational diabetes mellitus (GDM).
Boxplots show the distribution of EV-derived miRNAs in non-pregnant (NP), normal pregnant (NORM) and gestational diabetes (GDM) maternal plasma samples in (A) first trimester, (B) second trimester, (C) third trimester and (D) at delivery. Abundance is shown as Log-normalized read counts. Significance of the difference between groups was calculated using Kruskal–Wallis one-way analysis of variance followed by post-hoc analysis using Dunn’s test. Bonferroni corrected p-values demonstrate significant differences.
Fig 4.
Pregnancy specific EV-derived miRNAs at three trimesters of gestation in normal pregnancy.
Volcano plots show differential abundance of pregnancy-specific miRNAs from plasma EVs between non-pregnant (NP) and normal pregnancy (NORM) in (A) first, (B) second and (C) third trimesters of pregnancy and (D) at delivery. Filled circles represent no significant difference, triangles represent miRNA with upregulation and squares represent downregulation. Red denoting chromosome 19 microRNA clusters (C19MCs) while yellow denoting chromosome 14 microRNA clusters (C14MCs) and other non-C19MC or non-C14MC miRNAs are represented in blue.
Fig 5.
Differential abundance of EV-derived miRNAs in gestational diabetes mellitus (GDM) pregnancies.
Heatmap showing expression of placenta-specific C19MC and C14MC in EVs isolated from non-pregnant women (NP), and during 1st, 2nd, 3rd trimesters and at delivery (del) in normal pregnancy (NORM) and gestational diabetes (GDM) (A). Volcano plots display miRNAs from EVs isolated from gestational diabetes mellitus (GDM) (B) compared to normal pregnancies, during the first trimester of pregnancy. Red dots represent downregulated miRNAs, white dots represent miRNAs that are not significant for differential expression, blue dots represent upregulated miRNAs, green dots represent miRNAs from the C14MC and orange dots represent miRNAs from the C19MC.
Fig 6.
Differential abundance of miRNAs in placental tissue.
Volcano plots show expression of miRNAs in term placental tissue from pregnancies complicated by (A) GDM compared to normal pregnancies. Correlation between miRNAs from the term placenta and EV-derived miRNAs isolated from maternal plasma at delivery was performed in (B) normal pregnancy and (C) GDM.
Fig 7.
Predicted miRNA-target mRNA interaction in gestational diabetes mellitus (GDM) during early gestation.
A). The miRNA-Target mRNA network depicts targets of the top 20 differentially expressed EV-derived miRNAs from GDM in the first trimester of pregnancy. Blue triangles represent miRNAs while red dots refer to their targets. Dot plots represent functional enrichment analysis depicting target genes of selected miRNAs. The Y-axis displays the annotation categories KEGG pathways (B) and Reactome pathways (C) while the X-axis depicts the selected miRNAs.
Fig 8.
Receiver operator characteristics (ROC) curve for gestational diabetes GDM early gestation prediction.
(A&C) Receiver operator characteristic (ROC) curve generated using miRNAs isolated from plasma derived EVs of first trimester of pregnancy with GDM (n = 14) versus normal pregnancy (n = 7). (B&D) violin plot of predicted probability for GDM with 0 = low probability and 1 = high probability in normal (NORM) and GDM. (A,B = approach using all miRNAs detected in GDM, C,D = approach using differentially expressed genes alone).