Proteomic identification of biomarkers in maternal plasma that predict the outcome of rescue cerclage for cervical insufficiency

OBJECTIVE: We sought to identify plasma protein biomarkers that are predictive of the outcome of rescue cerclage in patients with cervical insufficiency. METHODS: This retrospective cohort study included 39 singleton pregnant women undergoing rescue cerclage for cervical insufficiency (17–25 weeks) who gave plasma samples. Three sets of pooled plasma samples from controls (cerclage success, n = 10) and cases (cerclage failure, n = 10, defined as spontaneous preterm delivery at <33 weeks) were labeled with 6-plex tandem mass tag (TMT) reagents and analyzed by liquid chromatography-tandem mass spectrometry. Differentially expressed proteins between the two groups were selected from the TMT-based quantitative analysis. Multiple reaction monitoring-mass spectrometry (MRM-MS) analysis was further used to verify the candidate proteins of interest in patients with cervical insufficiency in the final cohort (n = 39). RESULTS: From MRM-MS analysis of the 40 proteins showing statistically significant changes (P < 0.05) from the TMT-based quantitative analysis, plasma IGFBP-2, PSG4, and PGLYRP2 levels were found to be significantly increased, whereas plasma MET and LXN levels were significantly decreased in women with cerclage failure. Of these, IGFBP-2, PSG4, and LXN levels in plasma were independent of cervical dilatation. A multiple-biomarker panel was developed for the prediction of cerclage failure, using a stepwise regression procedure, which included the plasma IGFBP-2, PSG4, and LXN (area under the curve [AUC] = 0.916). The AUC for this multiple-biomarker panel was significantly greater than the AUC for any single biomarker included in the multi-biomarker model. CONCLUSIONS: Proteomic analysis identified useful and independent plasma biomarkers (IGFBP-2, PSG4, and LXN; verified by MRM) that predict poor pregnancy outcome following rescue cerclage. Their combined analysis in a multi-biomarker panel significantly improved predictability.