Using genome-wide transcriptome analysis by RNA sequencing of first trimester plasma RNA, we tested whether the identification of pregnancies at risk of developing pre-eclampsia with or without preterm birth or growth restriction is possible between weeks 9–14, prior to the appearance of clinical symptoms. We implemented a metaheuristic approach in the self-learning SVM algorithm for differential gene expression analysis of normal pregnancies (n = 108), affected pregnancies (n = 34) and non-pregnant controls (n = 19). Presymptomatic candidate markers for affected pregnancies were validated by RT-qPCR in first trimester samples (n = 34) from an independent cohort. PRKG1 was significantly downregulated in a subset of pregnancies with birth weights below the 10thpercentile as shared symptom. The NRIP1/ZEB2 ratio was found to be upregulated in pregnancies with pre-eclampsia or trisomy 21. Complementary quantitative analysis of both the linear and circular forms of NRIP1 permitted discrimination between pre-eclampsia and trisomy 21. Pre-eclamptic pregnancies showed an increase in linear NRIP1 compared to circular NRIP1, while trisomy 21 pregnancies did not.

Background In the first trimester of pregnancy, the maternal platelet is directly involved in a positive feedback mechanism that facilitates invasion of the extravillous trophoblast into the maternal spiral arteries. Dysfunctional trophoblast invasion with defective deep placentation is primordial in the etiology of the “great obstetrical syndromes.” Methods In this proof-of-concept study, using transcriptome analysis of circular RNA (circRNA) following RNA sequencing of maternal platelets, we tested whether pregnancy-specific circRNA markers could be identified in the first trimester of normal pregnancies. Differential transcript expression analysis of circRNAs, as predicted by Accurate CircRNA Finder Suite, CircRNA Identifier (version 2), and Known and Novel Isoform Explorer, was done using thromboSeq.R with variation of multiple settings. Test performance was checked for (a) de novo circRNA identification using the novel platelet-specific Plt-circR4 as a positive control, (b) complete segregation of groups (pregnant vs nonpregnant) after heat map–dendrogram clustering, (c) identification of pregnancy-specific circRNA markers at a false discovery rate (FDR) <0.05, and (d) confirmation of differentially expressed circRNA markers with an FDR <0.05 by an independent method, reverse transcription–quantitative PCR. Results Of the differentially expressed circRNAs with P values <0.05, 41 circRNAs were upregulated (logFC >2), and 52 circRNAs were downregulated (logFC less than −2) in first-trimester platelet RNA. Of these, nuclear receptor-interacting protein 1 circRNA covering exons 2 and 3 of the 5′-untranslated region was pregnancy specific with upregulation in first-trimester maternal platelets compared to nonpregnant controls. Conclusion CircRNA sequencing of first-trimester maternal platelets permits the identification of novel pregnancy-specific RNA biomarkers. Future use could include the assessment of maternal and fetal well-being.