Serum N-Glycosylation RPLC-FD-MS Assay to Assess Colorectal Cancer Surgical Interventions

Abstract

A newly developed analytical strategy was applied to profile the total serum N-glycome of 64 colorectal cancer (CRC) patients before and after surgical intervention. In this cohort, it was previously found that serum N-glycome alterations in CRC were associated with patient survival. Here, fluorescent labeling of serum N-glycans was applied using procainamide and followed by sialic acid derivatization specific for α2,6- and α2,3-linkage types via ethyl esterification and amidation, respectively. This strategy allowed efficient separation of specific positional isomers on reversed-phase liquid chromatography-fluorescence detection-mass spectrometry (RPLC-FD-MS) and complemented the previous glycomics data based on matrix-assisted laser desorption/ionization (MALDI)-MS that did not include such separations. The results from comparing pre-operative CRC to post-operative samples were in agreement with studies that identified a decrease in di-antennary structures with core fucosylation and an increase in sialylated tri- and tetra-antennary N-glycans in CRC patient sera. Pre-operative abundances of N-glycans showed good performance for the classification of adenocarcinoma and led to the revisit of the previous MALDI-MS dataset with regard to histological and clinical data. This strategy has the potential to monitor patient profiles before, during, and after clinical events such as treatment, therapy, or surgery and should also be further explored.

Keywords: N-glycosylation; colorectal cancer; mass spectrometry; reversed-phase liquid chromatography; serum; sialic acid derivatization; surgery.

Figure 1

Analysis of pre- and post-operative N-glycans in CRC by RPLC-FD-MS. Serum samples were taken pre- and post-operation from the same patient. Each sample underwent fluorescent labeling and sialic acid derivatization N-glycan workflow followed by RPLC-FD-MS measurement. Three quantification approaches were used for data processing, namely FD, MS, and a combination of these two approaches via FD-MS. Direct N-glycan and derived glycosylation traits were determined in order to perform data analysis and comparison of the two groups. Adapted with permission from Moran, A.B., et al., 2022. Analytical Chemistry, 94(18), pp. 663–6648 [16]. Copyright 2022 American Chemical Society.

Figure 2

Significant direct N-glycan traits. Eight significant direct N-glycan traits were found between pre-operative (pre-, n = 64) and post-operative (post-, n = 64) samples that were also previously validated [14]: H4N4F1, H5N4F1, H5N5F1, H5N4F1S_2,61, H5N4F1S_2,32, H6N5F1S_2,31S_2,61, H6N5F1S_2,31S_2,62 (isomer 1), and H6N5S_2,63. Asterisks denote significance with p-value < 0.05 (*), 0.01 (**), 0.001 (***). Abbreviations: hexose (H), N-acetylhexosamine (N), fucose (F), amidated α2,3-linked N-acetylneuraminic acid (Am), ethyl esterified α2,6-linked N-acetylneuraminic acid (E), and isomer (iso).

Figure 3

Significant derived glycosylation traits. Seven significant derived glycosylation traits were found between pre-operative (pre-, n = 64) and post-operative (post-, n = 64) samples that were also previously validated [14]: A2F (fucosylated diantennary glycans), AG (galactosylation per antenna), AS (sialic acid per antenna), A2S (sialic acids per diantennary glycan), A3FAm (α2,3-linked sialic acids per fucosylated triantennary glycan), AFE (α2,6-linked sialic acids per fucosylated glycan), and A2E (α2,6-linked sialic acids per diantennary glycan). Asterisks denote significance with p-value < 0.05 (*), 0.01 (**), 0.001 (***), 0.0001 (****). Monosaccharide ranges are specified via plus-minus (±; 0–4) and plus (+; 1–4) symbols. Abbreviations: antennae (A), fucose (F), galactosylation (G), N-acetylneuraminic acid (S), amidated α2,3-linked N-acetylneuraminic acid (Am), and ethyl esterified α2,6-linked N-acetylneuraminic acid (E).

Figure 4

Positional isomeric N-glycans that show significant differences in CRC. Four positional isomers with significant differences were observed, as denoted by each row. (A1–D1) Boxplots of relative abundances observed in pre- (n = 64) and post-operative (n = 64) samples for each isomeric form. (A2–D2) Fold-change in CRC (pre-/post-, n = 64) for each isomer whereby an increase (>1) or a decrease (<1) may be observed. Asterisks denote significance with p-value < 0.05 (*), 0.01 (**), 0.001 (***), 0.0001 (****). Abbreviations: Not significant (ns), hexose (H), N-acetylhexosamine (N), fucose (F), amidated α2,3-linked N-acetylneuraminic acid (Am), and ethyl esterified α2,6-linked N-acetylneuraminic acid (E).

Figure 5

Differences in histological type between pre- and post-operative CRC samples. Three N-glycans differentiated significantly based upon histological type, adenocarcinoma (n = 57), mucinous adenocarcinoma (n = 7), and signet-ring cell carcinoma (n = 3), as denoted by each row: row A (H5N4F1), row B (H4N4F1) and row C (H5N4F1S_2,61). (A1–C1) Boxplots of relative abundances observed in pre- and post-operative samples. (A2–C2) Trend observed following surgery, from pre- to post-operative samples from the same patients. The 95% CI is shown as colored bands around each line. Asterisks denote significance with p-value < 0.05 (*), 0.01 (**), 0.001 (***). Abbreviations: Not significant (ns), hexose (H), N-acetylhexosamine (N), fucose (F), amidated α2,3-linked N-acetylneuraminic acid (Am), and ethyl esterified α2,6-linked N-acetylneuraminic acid €.