Urinary proteomics for prediction of preeclampsia

Abstract

Preeclampsia is a major determinant of fetal and maternal morbidity and mortality. We used a proteomic strategy to identify urinary biomarkers that predict preeclampsia before the onset of disease. We prospectively collected urine samples from women throughout pregnancy. Samples from gestational weeks 12 to 16 (n=45), 20 (n=50), and 28 (n=18) from women who subsequently had preeclampsia develop were matched to controls (n=86, n=49, and n=17, respectively). We performed capillary electrophoresis online coupled to micro-time-of-flight mass spectrometry. Disease-specific peptide patterns were generated using support vector machine-based software. Candidate biomarkers were sequenced by liquid chromatography-tandem mass spectrometry. From comparison with nonpregnant controls, we defined a panel of 284 pregnancy-specific proteomic biomarkers. Subsequently, we developed a model of 50 biomarkers from specimens obtained at week 28 that was associated with future preeclampsia (classification factor in cases, 1.032 ± 0.411 vs controls, -1.038 ± 0.432; P<0.001). Classification factor increased markedly from week 12 to 16 to 28 in women who subsequently had preeclampsia develop (n=16; from -0.392 ± 0.383 to 1.070 ± 0.383; P<0.001) and decreased slightly in controls (n=16; from -0.647 ± 0.437 to -1.024 ± 0.433; P=0.043). Among the biomarkers are fibrinogen alpha chain, collagen alpha chain, and uromodulin fragments. The markers appear to predict preeclampsia at gestational week 28 with good confidence but not reliably at earlier time points (weeks 12-16 and 20). After prospective validation in other cohorts, these markers may contribute to better prediction, monitoring, and accurate diagnosis of preeclampsia.