HILIC-FLD-based characterization of maternal plasma N-glycosylation profiles as novel diagnostic and predictive biomarkers for hemolytic disease of the fetus and newborn

Abstract

ABO incompatibility-induced hemolytic disease of the fetus and newborn (HDFN) lacks effective preventive measures, necessitating advanced strategies for early prenatal prediction. This study investigated whether pregnancy-associated maternal plasma N-glycome alterations enable accurate prediction and diagnosis of ABO-mediated HDFN. We conducted ultra high-resolution plasma N-glycomics using hydrophilic interaction liquid chromatography with fluorescence detection (HILIC-FLD) in a discovery cohort (40 HDFN cases vs. matched controls). Glycomic predictive models were developed through logistic regression of chromatographic glycan peaks (LglycoP) and feature-level analysis of glycosylation traits (LglycoF). Model performance was validated in an independent cohort (n = 42) using area under the receiver operator curve (AUC-ROC) with bootstrap resampling for stability assessment. Significant N-glycome dysregulation was identified in HDFN pregnancies, with 5 discriminative glycan peaks and altered glycosylation patterns featured by branching, trisialylation, di- and tri-galactosylation of glycans. The LglycoP model demonstrated robust capability of prediction and diagnosis of in both the discovery and validation cohort (AUC > 0.80), respectively. This study establishes maternal plasma N-glycomic profiling as a precision tool for early prediction and diagnosis of ABO-HDFN, particularly for clinical false positive category of antibody titer test. Identification of unique plasma dysregulation of N-glycome features in ABO-HDFN pregnancies suggest potential mechanistic links between maternal glyco-inflammatory responses and fetal erythrocyte opsonization, offering new perspectives on ABO-HDFN pathogenesis.