Serum N-Glycans as Independent Predictors of Death: A Prospective Investigation in the AEGIS Cohort

Abstract

Total N-glycome in blood serum or plasma provides information about all serum/plasma protein enzymatic glycosylation, a tightly regulated cotranslational and post-translational modification. Total plasma/serum N-glycome has shown specific patterns (signatures) in patients with high-mortality pathologies, such as cancer and cardiovascular diseases; thus, we explored the capacity of total serum N-glycome to predict mortality in a general adult population. This prospective cohort study was performed in a municipality in Spain including a random sample of 1516 adults. Participants were profiled for total serum N-glycome at baseline. Serum enzymatic N-glycan release was performed on a robotic platform followed by hydrophilic interaction chromatography-ultraperformance liquid chromatography glycan separation. The computerized medical records were checked at a median follow-up of 7.52 years to collect the date and cause of all deaths. N-glycan groups from total serum were used to develop mortality prediction models. Total serum N-glycome peak (GP) 16, mainly composed of A2[3]BG1S[3]1, predisposed to all-cause mortality; GP 22, mainly composed of FA2G2S[6]1, protected from all-cause mortality. The balance between them predicted all-cause mortality incidence over time (area under the curve [AUC], 0.810 [0.773-0.847]). Similar results were obtained for cancer mortality, with GPs 16, 17, 22, and 23 (AUC, 0.786 [0.728-0.843]); and for cardiovascular mortality, with GPs 7 and 9 (AUC, 0.747 [0.645-0.850]). Their predictive powers had an independent and additive effect on classical prediction factors. The balances between specific GPs are independent predictors of all-cause, cancer, and cardiovascular mortality and could contribute significantly to improving prognostic tools.