Molecular signatures in cervicovaginal secretions associated with the onset of labor and neonatal infectious risk: fundamental investigation and development of detection prototypes

Olivia Petitcolas

09 December 2026

Thesis defence

Pratical info

14h - 23h
Amphitheater Luton, Faculty building
Research professionnals and doctors
Reduced mobility access

Under the supervision of Céline Méhats, team From Gametes To Birth (FGTB)

 

Abstract:

Preterm birth, defined as delivery before 37 weeks of gestation, is a major cause of neonatal morbidity and mortality. In women presenting with preterm labor (PTL), predicting delivery within 7 days remains challenging despite its major impact on obstetric management. The PrediMAP project aims to improve this prediction by combining cervicovaginal biomarkers with clinical data and to develop a rapid diagnostic device for use in obstetric emergency units. PrediMAP is based on a protein signature of imminent labor (MCP-1, CD14, CD163, CCL-18, CD30, IL33-R, CD66a, TGFα). Analysis of the first half of a prospective cohort of 800 women presenting with PTL (PrediMAP cohort 1) confirmed that three of these proteins were among the best-performing biomarkers, supporting the hypothesis of increased inflammatory activity at the onset of labor. However, the initial algorithm showed limited performance (AUC = 0.85), suggesting that this signature did not capture all cases of imminent preterm birth. Analysis of misclassified patients identified three additional differentially expressed proteins: IGFBP-1, VEGF and IL-1β. These markers may reflect distinct pathophysiological mechanisms related, respectively, to preterm premature rupture of membranes, cervical insufficiency and infection. These findings support a model in which sterile inflammation represents a predominant mechanism, while alternative pathways involving membrane rupture, cervical dysfunction or infection may account for other cases of preterm birth. The final combination includes three biomarkers and three cervical clinical parameters. In a test set of 20 patients (10 delivering preterm within 7 days and 10 after 7 days), this combination achieved an AUC of 0.96. A second aim of this thesis is the development of an innovative multiplex immuno-PCR kit to quantify biomarkers from a cervicovaginal swab using Chronos, the ultra-fast PCR thermocycler developed and marketed by BforCure. Immuno-PCR results will be combined with clinical data, and the algorithm will provide an individualized risk estimate for preterm delivery within 7 days in obstetric emergency units. Correlation between immuno-PCR and the reference method has been demonstrated. To improve robustness and secure sensitive steps, the protocol has been partially automated. The next steps will include measurement of the three biomarkers in cohort 1 samples, followed by validation in a new prospective cohort of 1,800 women, together with a multicenter usability and reproducibility study based on 200 runs. A third aim is to improve prediction of early-onset neonatal bacterial infection. Two biomarkers were identified in the cervicovaginal secretions of women whose newborns developed infection. Their combination with vaginal microbiota data achieved an AUC of 0.84. These findings support the development of a prototype combining biomarker quantification with molecular characterization of the vaginal microbiota by gel-based PCR on Chronos. Further development will focus on multiplexing six specific probes to characterize the vaginal microbiota and integrate these data into a predictive algorithm. Overall, this work illustrates a translational approach ranging from the identification of biological signatures to their integration into rapid diagnostic tools. 4 Ultimately, these strategies aim to improve personalized risk stratification for both preterm birth and early-onset neonatal bacterial infection in obstetric care.

Key words: Preterm birth, delivery, infection, biomarkers, microbiota, algorithm, immuno-PCR, prediction