Modeling of the Desialylated Human Serum N-glycome for Molecular Diagnostic Applications in Inflammatory and Malignant Lung Diseases

Background: Immunoglobulin G and A, transferrin, haptoglobin and alpha-1- antitrypsin represent approximately 85% of the human serum glycoproteome and their N-glycosylation analysis may lead to the discovery of important molecular disease markers. However, due to the labile nature of the sialic acid residues, the desialylated subset of the serum N-glycoproteome has been traditionally utilized for diagnostic applications.

Objective: Creating a five-protein model to deconstruct the overall N-glycosylation fingerprints in inflammatory and malignant lung diseases.

Methods: The N-glycan pool of human serum and the five high abundant serum glycoproteins were analyzed. Simultaneous endoglycosidase/sialidase digestion was followed by fluorophore labeling and separation by CE-LIF to establish the model. Pooled serum samples from patients with COPD, lung cancer (LC) and their comorbidity were all analyzed.

Results: Nine significant (>1%) asialo-N-glycan structures were identified both in human serum and the standard protein mixture. The core-fucosylated-agalacto-biantennary glycan differentiated COPD and LC and both from the control and the comorbidity groups. Decrease in the core-fucosylated-agalacto-biantennary-bisecting, monogalacto and bigalacto structures differentiated all disease groups from the control. The significant increase of the fucosylated-galactosylated-triantennary structure was highly specific for LC, to a medium extent for COPD and a lesser extent for comorbidity. Also, some increase in the afucosylated-galactosylated-biantennary structure in all three disease types and afucosylated-galactosylated-triantennary structures in COPD and LC were observed in comparison to the control group.

Conclusion: Our results suggested that changes in the desialylated human serum Nglycome hold glycoprotein specific molecular diagnostic potential for malignant and inflammatory lung diseases, which can be modeled with the five-protein mixture.