A Robust and Versatile Automated Glycoanalytical Technology for Serum Antibodies and Acute Phase Proteins: Ovarian Cancer Case Study

Abstract

The direct association of the genome, transcriptome, metabolome, lipidome and proteome with the serum glycome has revealed systems of interconnected cellular pathways. The exact roles of individual glycoproteomes in the context of disease have yet to be elucidated. In a move toward personalized medicine, it is now becoming critical to understand disease pathogenesis, and the traits, stages, phenotypes and molecular features that accompany it, as the disruption of a whole system. To this end, we have developed an innovative technology on an automated platform, "GlycoSeqCap," which combines N-glycosylation data from six glycoproteins using a single source of human serum. Specifically, we multiplexed and optimized a successive serial capture and glycoanalysis of six purified glycoproteins, immunoglobulin G (IgG), immunoglobulin M (IgM), immunoglobulin A (IgA), transferrin (Trf), haptoglobin (Hpt) and alpha-1-antitrypsin (A1AT), from 50 μl of human serum. We provide the most comprehensive and in-depth glycan analysis of individual glycoproteins in a single source of human serum to date. To demonstrate the technological application in the context of a disease model, we performed a pilot study in an ovarian cancer cohort (n = 34) using discrimination and classification analyses to identify aberrant glycosylation. In our sample cohort, we exhibit improved selectivity and specificity over the currently used biomarker for ovarian cancer, CA125, for early stage ovarian cancer. This technology will establish a new state-of-the-art strategy for the characterization of individual serum glycoproteomes as a diagnostic and monitoring tool which represents a major step toward understanding the changes that take place during disease.

Keywords: acute phase proteins; antibodies; glycomics; glycoproteins; glycosylation; ovarian cancer; plasma or serum analysis.

Graphical abstract

Fig. 1.

Multiplexed automated serial capture of glycoprotein N-glycoprofiling. 96-well format robotic platform (A), specific anti-glycoprotein capture resin packed in PhyNexus phytip (B), serial capture of selected glycoproteins 1.Trf, 2.IgG, 3.IgM, 4.IgA, 5.Hpt, 6.A1AT (C), 1D SDS-PAGE separation of multiplexed automated capture of six selected glycoproteins from pooled human serum using PhyNexus Phytips. Lane 1: Protein Marker, Lane 2: Trf Standard, Lane 3: Bound Trf, Lane 4: Bound IgG, Lane 5: Bound IgM, Lane 6: Bound IgA, Lane 7: A1AT Standard, Lane 8: Bound A1AT, Lane 9: Hpt Standard, Lane 10: Bound Hpt. The protein bands marked with arrows are traces of albumin protein (non-glycosylated) (D), automated glycoprotein sample preparation, PNGaseF release and aminoquinoline carbamate (AQC) labeling of N-glycans (E) and finally ultra-high performance liquid chromatography (UPLC) separation and glycan structural analysis (F).

Fig. 2.

UPLC-HILIC-FLD chromatograms of AQC labeled N-glycans released from human serum for affinity purified glycoproteins IgG, IgM and IgA, Trf, Hpt, A1AT and 2-AB labeled N-glycans released from total serum. Only major glycans are annotated. All major N-glycans identified within the total serum UPLC chromatogram are present in the individual glycoprotein chromatograms which showcases that serum glycosylation is largely dominated by the highest abundance proteins (IgG, IgM and IgA, Trf, Hpt and A1AT). SNFG glycan nomenclature is used throughout.

Fig. 3.

Serial capture and N-glycoprofiling of human serum IgG, IgM and IgA, purified from human serum. The 25 IgG glycan peak areas (G1-G25), 24 IgM peak areas (M1–24) and 25 IgA glycan peak areas (A1-A25) plotted for five technical replicates over three different days (3A). The standard error shown as error bars. Sequencing of AQC labeled IgG, IgM and IgA N-glycans visualized by UPLC-HILIC chromatograms using exoglycosidase enzymes with glucose units (GU) to facilitate glycan identification (3B). Digestion of AQC labeled IgG, IgM and IgA N-glycans with addition of sialidase (ABS), galactosidase (BTG), hexosaminidase (GUH), fucosidase (BKF) and mannosidase (JBM) in the following order ABS, ABS+BTG, ABS+BTG+GUH, ABS+BTG+GUH+BKF and ABS+BTG+GUH+BKF+JBM. For IgG, arrows indicate the cleavage of sugar residues for selected peaks: major glycans FA2G2S1 and FA2G2S2. For IgM, arrows indicate the cleavage of sugar residues for selected peaks: major glycans FA2G2S1 and FA2BG2S1. For IgA, arrows indicate the cleavage of sugar residues for selected peaks: major glycans A2G2S1, FA2G2S2 and FA2BG2S2. SNFG nomenclature is used for glycan representation.

Fig. 4.

Serial capture and N-glycoprofiling of human serum Trf, Hpt and A1AT, purified from human serum. The 28 Trf glycan peak areas (T1-T28), 31 Hpt glycan peak areas (H1-H31) and 28 A1AT glycan peak areas (AT1-AT28) were plotted for five technical replicates over three different days (4A). The standard error shown as error bars. Sequencing of AQC labeled Trf, Hpt and A1AT N-glycans visualized by UPLC-HILIC chromatograms using exoglycosidase enzymes with glucose units (GU) to facilitate glycan identification (4B). Digestion of AQC labeled Trf, Hpt and A1AT N-glycans with addition of sialidase (ABS), galactosidase (BTG), hexosaminidase (GUH), fucosidase (BKF) and mannosidase (JBM) in the following order ABS, ABS+BTG, ABS+BTG+GUH, ABS+BTG+GUH+BKF and ABS+BTG+GUH+BKF+JBM. In Trf, arrows indicate the cleavage of sugar residues for selected peaks: major glycans A2G2S(3,6)2 and A2G2S2(6,6)2. In Hpt, arrows indicate the cleavage of sugar residues for selected peaks: major glycans A3G3S3, A2G2S2 and A2G2S1. In A1AT, arrows indicate the cleavage of sugar residues for selected peaks: major glycans A3G3S3 and A2G2S2.SNFG nomenclature is used for glycan representation.

Fig. 5.

Glycosylation traits for selected antibodies (IgG, IgM, IgA) and acute phase proteins (Trf, Hpt, A1AT) for healthy human serum are presented (5A). The values are represented as a % of the total glycosylation and are calculated from data provided in supplemental Table S15. Glycosylation features of antibodies (IgG, IgM and IgA) and acute phase proteins (Trf, Hpt, A1AT) are presented as control charts (5B and 5C).

Fig. 6.

Statistically significant GPs and glycosylation traits for each glycoproteins and clinical parameters (Y = yes) with p values <0. 05 (corrected for multiple testing error using a 5% FDR proposed by Benjamini-Hochberg (40)) for ovarian cancer cohort of normal versus borderline, metastatic versus borderline and normal versus metastatic (6A). The major glycan for the each statistically significant GP is presented (6B). Boxplots are presented for the statistically significant GPs, glycosylation traits and logCA125 for borderline (red), metastatic (green) and normal (blue) samples (6C).

Fig. 7.

Discrimination performance for noteworthy glycosylation traits (Hpt) for ovarian cancer cohort of normal versus borderline, normal versus metastatic and borderline versus metastatic. Linear regression model including AUC, SEN and SPE (7A), cluster analysis using PCA using Hpt derived traits as input for normal (blue, n = 7) versus borderline (orange, n = 6) separation, normal (blue, n = 7) versus metastatic (orange, n = 18) and borderline (orange, n = 6) versus metastatic blue, n = 18) respectively the following (7B) to highlight part B. In brackets the variance of the principal component.