Data Independent Acquisition Mass Spectrometry Enhanced Personalized Glycosylation Profiling of Haptoglobin in Hepatocellular Carcinoma

Abstract

Aberrant glycosylation has gained significant interest for biomarker discovery. However, low detectability, complex glycan structures, and heterogeneity present challenges in glycoprotein assay development. Using haptoglobin (Hp) as a model, we developed an integrated platform combining functionalized magnetic nanoparticles and zwitterionic hydrophilic interaction liquid chromatography (ZIC-HILIC) for highly specific glycopeptide enrichment, followed by a data-independent acquisition (DIA) strategy to establish a deep cancer-specific Hp-glycosylation profile in hepatitis B virus (HBV, n = 5) and hepatocellular carcinoma (HCC, n = 5) patients. The DIA strategy established one of the deepest Hp-glycosylation landscapes (1029 glycopeptides, 130 glycans) across serum samples, including 54 glycopeptides exclusively detected in HCC patients. Additionally, single-shot DIA searches against a DIA-based spectral library outperformed the DDA approach by 2-3-fold glycopeptide coverage across patients. Among the four N-glycan sites on Hp (N-184, N-207, N-211, N-241), the total glycan type distribution revealed significantly enhanced detection of combined fucosylated-sialylated glycans, which were the most dominant glycoforms identified in HCC patients. Quantitation analysis revealed 48 glycopeptides significantly enriched in HCC (p < 0.05), including a hybrid monosialylated triantennary glycopeptide on the N-184 site with nearly none-to-all elevation to differentiate HCC from the HBV group (HCC/HBV ratio: 2462 ± 766, p < 0.05). In summary, DIA-MS presents an unbiased and comprehensive alternative for targeted glycoproteomics to guide discovery and validation of glyco-biomarkers.

Keywords: data-independent acquisition (DIA); glycoprotein; haptoglobin; hepatocellular carcinoma (HCC).

Figure 1

Schematic diagram of the integrated platform for identification and quantitation of haptoglobin glycopeptide using DIA-MS. By combining hemoglobin-conjugated magnetic nanoparticles (Hb@MNPs) specific to haptoglobin with a well-established ZIC-HILIC separation strategy, we expect highly sensitive and specific haptoglobin glycopeptide enrichment. We established an Hp-specific glycopeptide spectral library using DIA data from pooled serum samples, integrating Byonic and Skyline software. Individual patient Hp was analyzed via single-shot DIA spiked with iRT for N-glycopeptide identification and label-free quantitation in the HCC and HBV groups. Statistical analysis evaluated differentially expressed site-specific N-glycopeptides between these cohorts.

Figure 2

Construction and comparative analysis of spectral libraries generated using DDA and DIA data sets obtained from triplicate analysis of the HCC serum sample. (A) Comparison of the number of MS/MS spectrum containing oxonium ions from glycopeptide (GP) and sialylated glycopeptide (SG) between DDA and DIA methods. (B) Total number of precursors and glycopeptides included in the 3 spectral libraries constructed by DDA, DIA, or hybrid DDA/DIA data sets. (C) Glycan type distribution from glycopeptides commonly present in the DDA and DIA libraries or unique glycopeptides in the DIA spectral library.

Figure 3

Representative MS/MS spectra of identified Hp glycopeptide from HCC serum. Comparison of 2 identical trisialyl-triantennary glycans (A) from DDA (m/z 1552.34, z = 3+, Byonic score = 673.4) and (B) from DIA (m/z = 1553.0088, z = 3+, Byonic score = 939.5). Blue squares, GlcNAc; green circles, Man; yellow circles, Gal; purple diamond, sialic acid.

Figure 4

Comparison of single-shot DDA and DIA analysis using 3 technical replicates of the HCC serum sample. (A) Comparison of identified Hp glycopeptide from each glycosite between the DDA and library-based DIA methods. (B) Venn diagram of overlapping glycopeptides identified from the DDA, DIA, and hybrid DDA/DIA libraries. The glycan type distribution from the uniquely identified glycopeptide from the DIA library (217 glycopeptides) was shown to demonstrate the predominant identification of complex types of glycan from library-based DIA. (C) The abundance distribution of all glycopeptides and (D) comparison of common and unique glycopeptide abundance between the DDA and DIA methods.

Figure 5

Comparison of DDA and DIA identification result for individual patient sera. Identification of unique Hp glycopeptides (A) per N-glycosylation site and (B) per glycan type distribution from individual HCC (n = 5) and HBV (n = 5) patients. (C) Venn diagram of overlapped glycopeptide from DDA and DIA across each disease group. The glycan type distribution from 54 glycopeptides identified in the DIA method.

Figure 6

Quantification and differential abundance of glycopeptide identified from HCC and HBV patients using the library-based DIA method. (A) Heatmap of differential Hp glycopeptide precursor and corresponding glycans (shown by N-glycosite, glycan composition, peptide modification, and m/z) categorized by hierarchical clustering. Dea: Deamidation, Oxi: Oxidation. (B) Representative figure of glycan structures of glycopeptides enriched in either HCC group (n = 5) or HBV group (n = 5) in each Hp glycosite (N184, N207, N211, and N241). The hollow rectangle with the dashed line categorizes glycans into 5 groups (yellow: fucosyl and sialyl; purple: sialyl; red: fucosyl; green: high mannose; blue: other glycans). The numbers indicate the total counts of categorized glycan structures in each glycan type for the corresponding glycosite.