First trimester circulating miR-208b-3p and miR-26a-1-3p are relevant to the prediction of gestational hypertension

Abstract

Background: Gestational hypertension (GH) is linked to an increased risk of cardiometabolic diseases for both mother and child, but we lack reliable biomarkers to identify high-risk women early in pregnancy. MicroRNAs (miRNAs) are small non-coding RNA that have emerged as promising biomarkers for pregnancy complications. We thus aimed to identify first trimester circulating miRNAs associated with GH and to build a miRNA-based algorithm to predict GH incidence.

Methods: We quantified miRNAs using next-generation sequencing in plasma samples collected at first trimester of pregnancy in Gen3G (N = 413, including 28 GH cases) and 3D (N = 281, including 21 GH cases) prospective birth cohorts. MiRNAs associated with GH in Gen3G (identified using DESeq2, p-value < 0.05) and replicated in 3D were included in a stepwise logistic regression model to estimate the probability of developing GH based on the miRNAs (normalized z-score counts) and maternal characteristics that contribute most to the model.

Results: We identified 28 miRNAs associated with the onset of GH later in pregnancy (p < 0.05) in the Gen3G cohort. Among these, three were replicated in the 3D cohort (similar fold change and p < 0.1) and were included in stepwise logistic regression models with GH-related risk factors. When combined with first trimester mean arterial pressure (MAP), miR-208b-3p and miR-26a-1-3p achieve an AUC of 0.803 (95%CI: 0.512-0.895) in Gen3G and 0.709 (95%CI: 0.588-0.829) in 3D. The addition of miR-208b-3p, and miR-26a-1-3p to the model significantly improves the prediction performance over that of MAP alone (p = 0.03). We then proposed low and high-risk thresholds, which could help identify women at very low risk of GH and those who could benefit from prevention monitoring throughout their pregnancy.

Conclusion: The combination of circulating miR-208b-3p and miR-26a-1-3p with first trimester MAP offers good performance as early predictors of GH. Interestingly, these miRNAs target pathways related to the cardiovascular system and could thus be relevant to the pathophysiology of GH. These miRNAs thus provide a novel avenue to identify women at risk and could lead to even more adequate obstetrical care to reduce the risk of complications associated with GH.

Keywords: Human diseases; MicroRNA (miRNA); Non-coding RNAs; Non-invasive biomarkers; Pregnancy-induced hypertension; Ribo-hormones.

Fig. 1

Volcano plot of the 2 170 plasma miRNAs detected by next-generation sequencing. The volcano plot shows the Log2(Fold Change) on the x-axis while the y-axis shows the significance (nominal p-value) as -Log₁₀(p-value). The vertical dotted lines represent a fold change of 20% while the horizontal dotted line corresponds to a nominal p-value = 0.05. The 20 miRNAs less abundant in women who develop GH are represented in orange while the 8 miRNAs more abundant in women who develop GH are in blue. The three replicated miRNAs in 3D are labelled on the plot. The volcano plot was generated using EnhancedVolcano package in R (35)

Fig. 2

Performance of GH prediction models by Receiver Operating Characteristics (ROC) curves. (A) ROC curves for each model in the Gen3G cohort (N = 384 normotensives and 28 GH),(B) ROC curves for each model in the 3D cohort (N = 249 normotensives and 20 GH). Abbreviations: AUC: area under the curve, CI: confidence interval, MAP: mean arterial pressure at v1 (visit 1, first trimester)

Fig. 3

Stratification of women according to their risk of developing GH based on predicted probabilities obtained from the logistic regression model incorporating miRNAs and mean arterial pressure. Each dot represents one woman, corresponding to the number of observations on the x-axis. The predicted probabilities resulting from the logistic regression model is shown on the y-axis. The proposed high and low-risk thresholds are represented by the solid and dotted lines, respectively. (A) Predicted probability of GH for each woman in the Gen3G cohort; (B) Predicted probability of GH for each woman in the 3D cohort