Identification of novel cell-free RNAs in maternal plasma as preterm biomarkers in combination with placental RNA profiles

• Published in: Journal of translational medicine 2023-04, Vol.21 (1), p.256-256, Article 256
• Main Authors: Jin, Heyue, Zhang, Yimin, Fan, Zhigang, Wang, Xianyan, Rui, Chen, Xing, Shaozhen, Dong, Hongmei, Wang, Qunan, Tao, Fangbiao, Zhu, Yumin
• Format: Article
• Language: English
• Subjects: Bioinformatics; Biomarkers; Biomarkers - metabolism; Cell-free RNA; Datasets; Female; Fetuses; Genes; Genomics; Humans; Infant, Newborn; Machine learning; Maternal & child health; Molecular modelling; Placenta; Placenta - metabolism; Plasma; Prediction; Pregnancy; Premature birth; Premature Birth - genetics; Premature Birth - metabolism; Preterm birth; Proteins; Ribonucleic acid; RNA; RNA - genetics; RNA - metabolism; Software; Transcriptome; Transcriptomes
• ISSN: 1479-5876; 1479-5876
• DOI: 10.1186/s12967-023-04083-w

Preterm birth (PTB) is the main driver of newborn deaths. The identification of pregnancies at risk of PTB remains challenging, as the incomplete understanding of molecular mechanisms associated with PTB. Although several transcriptome studies have been done on the placenta and plasma from PTB women, a comprehensive description of the RNA profiles from plasma and placenta associated with PTB remains lacking. Candidate markers with consistent trends in the placenta and plasma were identified by implementing differential expression analysis using placental tissue and maternal plasma RNA-seq datasets, and then validated by RT-qPCR in an independent cohort. In combination with bioinformatics analysis tools, we set up two protein-protein interaction networks of the significant PTB-related modules. The support vector machine (SVM) model was used to verify the prediction potential of cell free RNAs (cfRNAs) in plasma for PTB and late PTB. We identified 15 genes with consistent regulatory trends in placenta and plasma of PTB while the full term birth (FTB) acts as a control. Subsequently, we verified seven cfRNAs in an independent cohort by RT-qPCR in maternal plasma. The cfRNA ARHGEF28 showed consistence in the experimental validation and performed excellently in prediction of PTB in the model. The AUC achieved 0.990 for whole PTB and 0.986 for late PTB. In a comparison of PTB versus FTB, the combined investigation of placental and plasma RNA profiles has shown a further understanding of the mechanism of PTB. Then, the cfRNA identified has the capacity of predicting whole PTB and late PTB.