511: New-onset postpartum preeclampsia: Mechanisms and Prediction

Abstract

molecular signatures that will differentiate among various clinical phenotypes of HELLP and provide an accurate differential diagnosis of this syndrome from HELLP imitators. STUDY DESIGN: We analyzed urine samples from 6 groups of patients classified phenotypically as: non-hypertensive healthy controls (CRL, n¼9, 30AE1 wks), women with idiopathic proteinuria (IP, n¼15, 28AE2 wks), PE mild features (mPE, n¼16, 33AE1 wks), superimposed PE (spPE, n¼16, 31AE1 wks), PE severe features (sPE, n¼16, 32AE1 wks) and HELLP (n¼16, 29AE1 wks). In the discovery phase, urine proteins were enriched by CR precipitation and subjected to UPLC-tandem mass spectrometry. The data were analyzed with Myrimatch and IDPicker. Theoretical masses were calculated from the human Uniprot sequence database. Protein IDs from all samples were dimensionally reduced using multi-dimensional scaling. HELLP samples were clustered (Euclidean distance, average linkage) using silhouette score to control clustering quality. Characteristic proteins for the clusters were determined by variance analysis. RESULTS: 1,747 IDs were identified in the CR precipitate and merged into 690 protein groups. The false discovery rate was 1.12%. These IDs generated good separation for 3 major clinical phenotypes (Figure). Variance analysis yielded Immunoglobulin Kappa Locus protein and Immunoglobulin-like protein (role: immunoglobulin synthesis) for cluster 1, Keratin types I and II for cluster 2 (role: structural framework of epithelial cells), and three Alpha-1-Antitrypsin (A0A024R6I7, Q3I0J7, A7L8C5) variants (role: endogenous protease inhibitor) for cluster 3. CONCLUSION: We provide evidence that HELLP syndrome involves 3 subclinical phenotypes that can be differentiated through examination of the urinary misfoldome.