Every year, one in 13 babies are born prematurely in the UK. During pregnancy, the reproductive tract and cervico-vaginal fluid (CVF) is proposed to reflect inflammatory events which later contribute to early parturition. Exosomes, a specific subtype of secreted vesicles that have the capacity to transport bioactive molecules like miRNAs and proteins, are released from a wide range of cells including placental cells (Adam et al., 2017; Salomon et al., 2018). As the content of exosomes is cell type specific, we hypothesize that there will be distinct exosomal “fingerprints” associated with the risk of spontaneous preterm birth (sPTB). The aim of this study was to profile exosomes isolated from CVF obtained from women in the first and second trimesters of pregnancy who later had sPTB or term deliveries. Thirty-seven pregnant women (n=10 PTB <37 weeks; n=27 >37 weeks’ gestation) recruited as part of a larger prospective cohort study (INSIGHT REC No. 13/LO/0393; written informed consent) provided CVF between 10-23+6 weeks of pregnancy and clinical metadata. Exosomes were isolated from CVF by differential centrifugation and characterised by nanoparticle tracking analysis using the NanoSight NS500. An Illumina TrueSeq Small RNA kit was used to construct a small RNA library from CVF exosomal RNA and sequenced. Real time PCR was used to validate miRNA candidates in all samples. The exosomal proteomic profiles were identified in all samples by liquid chromatography mass spectrometry and quantified by Sequential Window Acquisition of All Theoretical (SWATH) analysis. LogitBoost regression was used to develop a classification model for early detection of sPTB cases using STATA software. Exosomes were identified as small vesicles (50-150 nm) in 30 CVF samples. Exosomal miRNA analysis revealed seven downregulated and nine upregulated miRNAs in sPTB cases compared with term. With a false discovery rate of 1 %, a total of 206 proteins within CVF exosomes were identified and quantified using SWATH. We identified three proteins that were significantly lower and one protein higher (p<0.05) in exosomes from sPTB compared with term. Using the miRNA and proteomic profiles of exosomes, a binomial classification algorithm was generated using a boosted logistic regression analysis (WEKA machine learning software). The algorithm was built using four miRNAs- and four proteins-associated with exosomes. The model delivered discrimination between women who had sPTB and term pregnancies, with an area under the curve of 0.822. We provide evidence that first and second trimester CVF exosomes contain measurable miRNA and protein signals. In this small pilot study, the CVF proteomic signal alone or in combination with miRNA shows potential in relation to the prediction of preterm birth. Validation of this novel exosomal biomarker panel is needed in a larger cohort of term and preterm pregnancies.

We thank the CRN funded research midwives and research assistants involved with the study, Action Medical Research and Borne Charity (Grant GN666), Rosetrees Trust, NIHR and the Tommy’s Charity.