Quantitative analysis of core fucosylation of serum proteins in liver diseases by LC-MS-MRM

Abstract

Aberrant core fucosylation of proteins has been linked to liver diseases. In this study, we carried out multiple reaction monitoring (MRM) quantification of core fucosylated N-glycopeptides of serum proteins partially deglycosylated by a combination of endoglycosidases (endoF1, endoF2, and endoF3). To minimize variability associated with the preparatory steps, the analysis was performed without enrichment of glycopeptides or fractionation of serum besides the nanoRP chromatography. Specifically, we quantified core fucosylation of 22 N-glycopeptides derived from 17 proteins together with protein abundance of these glycoproteins in a cohort of 45 participants (15 disease-free control, 15 fibrosis and 15 cirrhosis patients) using a multiplex nanoUPLC-MSMRM workflow. We find increased core fucosylation of 5 glycopeptides at the stage of liver fibrosis (i.e., N630 of serotransferrin, N107 of alpha-1-antitrypsin, N253 of plasma protease C1 inhibitor, N397 of ceruloplasmin, and N86 of vitronectin), increase of additional 6 glycopeptides at the stage of cirrhosis (i.e., N138 and N762 of ceruloplasmin, N354 of clusterin, N187 of hemopexin, N71 of immunoglobulin J chain, and N127 of lumican), while the degree of core fucosylation of 10 glycopeptides did not change. Interestingly, although we observe an increase in the core fucosylation at N86 of vitronectin in liver fibrosis, core fucosylation decreases on the N169 glycopeptide of the same protein. Our results demonstrate that the changes in core fucosylation are protein and site specific during the progression of fibrotic liver disease and independent of the changes in the quantity of N-glycoproteins. It is expected that the fully optimized multiplex LC-MS-MRM assay of core fucosylated glycopeptides will be useful for the serologic assessment of the fibrosis of liver.

Biological significance:

We have quantified the difference in core fucosylation among three comparison groups (healthy control, fibrosis and cirrhosis patients) using a sensitive and selective LC-MS-MRM method. Despite an overall increase in core fucosylation of many of the glycoproteins that we examined, core fucosylation changed in a protein- and site-specific manner. Moreover, increased and decreased fucosylation was observed on different N-glycopeptides of the same protein. Altered core fucosylation of N-glycopeptides might be used as an alternative serologic assay for the evaluation of fibrotic liver disease.

1. Introduction

N-glycosylation exerts critical roles in numerous physiological and pathological processes [1]. Core fucosylation is formed by the α−1,6-fucosyltransferase (Fut8) which transfers the fucosyl moiety from GDP-fucose to the innermost GlcNAc residue of N-linked oligosaccharides of glycoprotein [2–4]. Altered core fucosylation (α−1,6) is associated with several diseases [5,6] including gastric tumors [7] and light chain multiple myeloma [8]. An increased degree of core fucosylation has been observed in the progression of lung cancer [9], prostate cancer [10], breast cancer [11], and melanoma [12]. Notably, an overall increased core fucosylation has been associated with hepatocellular carcinoma (HCC) [13–16] and these changes can be observed in serum owing to the homeostatic function of liver secreted glycoproteins [17]. This is already explored in clinical assessment of alpha-fetoprotein L3 (AFP-L3), a core fucosylated form of alpha-fetoprotein (AFP) approved by the US FDA as a biomarker of the risk of HCC [18]. However, AFP-L3 is marginally effective at the early diagnosis of HCC [19] even though hepatocarcinogenesis was significantly suppressed in a Fut8 knock out animal model [20].

Wide distribution of core fucosylated N-glycans in human proteins was documented by improved mass spectrometric methods and it was suggested that specificity of core fucosylation depends on solvent accessibility of the N-glycosylation sites [21]. Due to the complexity of the heterogeneous N-glycoforms and their sub-stoichiometric representation, immunodepletion and further enrichment are often required for the characterization of core fucosylated N-glycoproteins in complex matrices like serum and plasma. High coverage of core fucosylated N-glycopeptides was achieved by combining lectin-based enrichment of core fucosylated proteins/peptides (e.g. Lens culinaris Agglutinin), followed by cleavage of glycans by endoglycosidase F3 (endoF3) and/or fractionation prior to the mass spectrometric analysis [22–28]. However, many of these studies were not conducted in a quantitative manner. For site-specific analysis of core fucosylation, stable isotopic labeling with iTRAQ [26] or TMT [28] was adopted in workflows combining enrichment with tandem mass spectrometry. These workflows were applied to the comparison of core fucosylation in prostate cancer cell lines [28] which identified 973 fucosylated glycopeptides and serum of patients with hepatocellular carcinoma (HCC) [26] which identified 1300 glycopeptides. Comparison of HCC with liver cirrhosis suggested increased fucosylation of 12 glycopeptides in HCC together with 14 glycopeptides increased in cirrhosis [26]. These complex workflows are, however, not directly applicable to clinical samples and a simplified assay comparing intensity of core fucosylated to non fucosylated peptide of isolated ceruloplasmin by nanoLC-LTQMS was published subsequently [23]. In this study, we present a simplified multiplex workflow that combines the endoF digest of glycopeptides with multiple reaction monitoring mass spectrometry (MRMMS), a well-developed targeted quantification workflow [29–36]. The targeted approach is applicable to the glycopeptides simplified by endoF digests because of the specificity and sensitivity of the glyco-peptide transitions in this setting [32,36]. These assays do not require enrichment at the protein or glycopeptide level and allow multiplex quantification of the core fucosylation directly in serum.

We applied the workflow to the analysis of serum samples from a set of control, fibrotic, and cirrhotic patients with the aim to establish a reliable method for the quantification of changes in core fucosylation associated with liver disease progression. Our results demonstrate that this LC-MS-MRM workflow enables simple, sensitive and selective quantification of core fucosylation and reveals protein- and site-specific changes in core fucosylation in the fibrotic liver disease.

2. Materials and methods

2.1. Chemicals and reagents

Ammonium bicarbonate, _DL-dithiothreitol (DTT), iodoacetamide (IAA), and endoglycosidases (endo) F1, F2, F3 were supplied by Sigma-Aldrich (St. Louis, MO, USA). Sequencing grade trypsin was purchased from Promega (Madison, WI, USA). 0.1% formic acid in ACN and 0.1% formic acid in H₂O were used as supplied from Honeywell (Philadelphia, PA, USA). LC/MS grade trifluoroacetic acid (TFA) and formic acid (FA) were from Thermo Fisher Scientific (Waltham, MA, USA). Water was purified and deionized with a Milli-Q system manufactured by Millipore (Bedford, MA, USA).

2.2. Sample information

Serum samples of 15 disease-free controls were collected at Georgetown University Medical Center under protocols approved by the Institutional Review Board. De-identified serum samples of 30 patients with various degree of liver fibrosis were obtained from the hepatitis C antiviral long-term treatment against cirrhosis trial (HALT-C), as described in our recent study [35]. The HALT-C trial, a prospective randomized controlled trial of 1050 patients, evaluated effect of long-term low-dose peginterferon alpha-2a (IFN) in participants that failed initial anti-HCV therapy with interferon [37,38]. Liver disease status of the HALT-C participants was classified into fibrosis (Ishak score 3–4) or cirrhosis (Ishak score 5–6) based on biopsy-evaluation. All the participants selected for this study were from the control arm of the HALTC trial that did not receive IFN treatment [35]. Method optimization was carried out on a sample pooled from the 45 serum samples of all participants (equal volume of each) and the pooled sample served as a quality control (QC) that was analyzed in parallel with the analyses of the 45 study samples. All the 45 samples were processed in parallel and analyzed individually as described below.

2.3. Sample digestion and deglycosylation

Serum samples (2 μL) were diluted in 140 μL of 50 mM NH₄HCO₃, 5 mM DTT was added and the reaction was incubated at 60 °C for 1 h. After cooling down to room temperature, 15 mM iodoacetamide was added for 30 min in the dark. Residual iodoacetamide was quenched with another aliquot of 5 mM DTT at room temperature for 20 min. Trypsin was added to the samples (1/40 (w/w)) and incubated at 37 °C overnight. After heating at 95 °C for 10 min, samples were dried down in a vacuum concentrator (Labconco). Deglycosylation with endoF1–F3 was performed according to the manufacturer’s protocol, with slight modifications. Specifically, samples were dissolved in 50 mM sodium acetate (pH 4.5), and then treated with 1 μL of endoF1 (Sigma E9762–1UN), 1 μL of endoF2 (Sigma E0639-.2UN) and 1 μL of endoF3 (Sigma E2264–.2UN) as supplied and 37 μL H₂O at 37 °C overnight. After heating at 95 °C for 10 min, the resulting products were dried down and dissolved in 0.1% formic acid for mass spectrometric analysis.

2.4. LC-MS/MS and LC-MS-MRM

To select core fucosylated (HexNAcFuc), non-fucosylated (HexNAc) peptides and unique peptides of N-glycoproteins for the LC-MRM-MS assays, tryptic digests of serum were analyzed on a NanoAcquity UPLC (Waters) coupled with a TripleTOF 6600 mass spectrometer (Sciex). Specifically, 1 μL of each sample prepared above was loaded onto a C18 Trap column (Waters Acquity UPLC Symmetry C18 NanoAcquity 10 K 2G V/M, 100 A, 5 μm, 180 μm × 20 mm) at 15 μL/min for 2 min. Peptides were then separated on an analytical column (Waters Acquity UPLC M-Class, peptide BEH C18 column, 300 A, 1.7 μm, 75 μm × 150 mm) at 40 °C. The flow rate was set at 400 nL/min with a 90-min gradient of buffer A (2% ACN, 0.1% formic acid) and buffer B (0.1% formic acid in ACN) was used for separation: 1% buffer B at 0 min, 5% buffer B at 1 min, 40% buffer B at 60 min, 99% buffer B at 65 min, 99% buffer B at 70 min. The gradient went back to 1% buffer B at 70.1 min and kept for 20 min for column equilibration. Mass spectral acquisition was operated with Analyst TF 1.7.1 software; each cycle consisted of a full scan (m/z 400–1600) and fifty information dependent acquisitions (IDAs) (m/z 100–1600). The dynamic exclusion time was set to 6 s, and 150 counts threshold for two repeated precursors. Collision energy was set automatically according to charge state and m/z of the precursor ion.

For quantification of the glycosylated and unique peptides in serum samples, tryptic digests were analyzed on a NanoAcquity UPLC (Waters) coupled with a QTRAP 6500 mass spectrometer (Sciex). The NanoAcquity UPLC (Waters) and the settings (e.g., gradient and flow rate) were the same as described above. Peptide and glycopeptide quantification on the QTRAP 6500 was carried out in the MRM mode with the ion spray voltage set at 2300 V, curtain gas 11, ion source gas 1 30, and the interface heater temperature 150 °C. The entrance potential (EP) was set as 10 V and collision cell exit potential (CXP) was 13 V. Q1 and Q3 were set at open and unit resolution, respectively. The MRM transitions used for monitoring of modified peptides and unique peptides are listed in Supplementary Tables S1 and S2. Collision energy (CE) for each MRM transition was optimized by a 5 V step optimization followed by a 2 V step fine tuning. Instrument control and data acquisition were performed by Sciex Analyst software (version 1.6.2).

2.5. Data processing and statistical analysis

Raw data files from the TripleTOF 6600 were processed with the Protein Pilot version 5.0 software (Sciex) utilizing the Paragon and Progroup algorithms and the integrated false discovery rate (FDR) analysis function. MS/MS data were searched against the NCBI Homo sapiens of the Uniprot-Sprot database downloaded on June 2, 2015. Search parameters are set as cys alkylation (Iodoacetamide), digestion (Trypsin), instrument (TripleTOF 6600), special factors (HexNAc emphasis and HexNAcFuc emphasis), species (Homo sapiens), ID Focus (Biological modifications), database (uniprot_sprot.fasta), search effort (Thorough), false discovery rate (FDR) analysis (Yes), and user modified parameter files (No). The proteins were inferred based on the ProGroup™ algorithm associated with the ProteinPilot software. The detected protein threshold in the software was set to the value which corresponded to 1% FDR. Peptides were defined as redundant if they had identical cleavage site(s), amino acid sequence, and modification. The mass spectra corresponding to HexNAcFuc/HexNAc-modified peptides were verified by spectral inspection.

LC-MS-MRM data from QTRAP 6500 was analyzed with MultiQuant 3.0 (Sciex) with manual confirmation; transitions with S/N > 10 were adopted for the quantification purposes. Peak intensities were used for peptide and glycopeptide quantification and data normalization. The MRM transition of the highest intensity for each peptide or glycopeptide was selected for quantification, while the remaining transitions were used to assure specificity of the targeted analyses. The core fucosylation level in each peptide was obtained by normalizing the intensity of HexNAcFuc-form peptide to the sum of intensities of both the HexNAcFuc-form and HexNAc-form. Duplicate measurements of each sample were averaged for quantitation.

Statistical analyses were performed with SAS release 9.4 (SAS Institute, Cary, NC). One-way Analysis of Variance (ANOVA) was used to evaluate the difference in core fucosylation among three comparison groups (healthy control, fibrosis and cirrhosis patients). In order to determine which pairs of groups are different after we reject H₀ from ANOVA, Bonferroni’s procedure at level 0.05 was performed for multiple testing. Box plots were used to show the distribution of core fucosylation percentage in three groups; all reported p-values are two sided. The same analysis was performed to determine the peak area of unique peptides of core fucosylated N-glycoproteins.

3. Results and discussion

We applied a glycosidase-assisted MRM workflow to the quantification of core fucosylation of serum proteins in fibrotic liver disease. The optimized method allows quantification of the glycosylated peptides without prior fractionation of serum or enrichment of the glycopeptides. In brief, discovery mode proteomics analysis was used to identify core fucosylated proteins and to select tryptic glycopeptides treated with endoglycosidases for the quantification of core fucosylation. This was followed by optimization of the LC-MS-MRM workflow used for multiplex quantification of 17 proteins and core fucosylation of 22 N-glycopeptides in 45 participants at progressing stages of fibrotic liver disease.

3.1. Glycopeptide selection and MRM optimization

In order to analyze core fucosylation, tryptic digests of serum proteins were treated with a combination of endoF1, endoF2, and endoF3. The endoF treatment results in truncated peptide-pairs with HexNAcFuc or HexNAc attached to the asparagine in the NXS/T sequon, as described previously [39–41]. Ratio of these glycoforms serves as a measure of the degree of core fucosylation of the proteins at the specific sites of glycan attachment. For optimal cleavage of the glycans, we optimized the conditions for endoF deglycosylation by testing different endoglycosidases, reaction times, and their combinations. In contrast to previous studies using endoF3 [23–27], we add endoF1 and endoF2 in order to improve the yield of products from different types of glycoforms. EndoF1 cleaves more efficiently oligomannose and hybrid structures, endoF2 prefers biantennary complex-type oligosaccharides, while endoF3 cleaves biantennary and triantennary glycans [39–41].Since the presence of core fucose increases the cleavage performance of endoglycosidases [42,43], we optimized the conditions for the cleavage of the non-fucosylated glycoforms of the peptides which were cleaved with > 90% efficiency for both bi- and tri-antennary structures. Limitation of this method is the known lack of cleavage of tetra-antennary structures by the endoF digests and core fucosylation of these glycoforms is not quantified. However, the tetra-antennary glycoforms generally account for a small percentage of all glycoforms (< 10%) and are not expected to affect substantially the glycopeptides in this study [44,45].

We evaluated 24 glycopeptide pairs (HexNAc and HexNAcFuc forms of the same glycopeptide) identified by LC-MS/MS via data dependent acquisition (Supplementary Tables S1 and S2). Retention time (RT) and fragmentation of each glycopeptide were used to select at least 3 transitions for the quantification of each glycoform with the following criteria: 1) the m/z of product ion usually greater than precursor m/z using singly or doubly charged y ions for the quantification, 2) b ions and transitions close to precursors were avoided to maximize selectivity, and 3) transitions with highest intensity were selected. Of the 24 pairs, 2 were excluded from the final multiplex quantification because one of the glycoforms (HexNAc and HexNAcFuc) was not reliably quantifiable in serum samples of > 20% of the participants; typical reason is the low degree of core fucosylation of the glycopeptide (Supplementary Tables S2).

Simultaneously, we included in the multiplex assays non-glycosylated peptides of the 17 proteins to assess changes in the protein abundance that accompany the changes in core fucosylation of these proteins in liver disease. Uniqueness of each selected peptide for the 17 proteins was confirmed by searching against UniProtKB/Swiss-Prot human protein database and Protein Basic Local Alignment Search Tool (BLAST) (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Candidate peptides were further filtered using the criteria listed below to reduce potential sources of variability: 1) peptides containing < 6 amino acids (aa) were avoided to ensure uniqueness, 2) large peptides (> 25 aa), potential ragged end peptides, and peptides with known post-translational modifications were excluded, 3) peptides containing missed tryptic cleavage sites (e.g. internal lysine or arginine residues) were avoided. LC-MS-MRM instrument parameters including collision energies (CE) were optimized to yield the highest sensitivity and linearity of response for all transitions. The final multiplex assay consists of the quantification of 22 pairs (HexNAc and HexNAcFuc forms) of glycopeptides and 17 unique peptides for the corresponding glycoproteins (Supplementary Tables S1).

3.2. MRM-based quantification of core fucosylation in liver disease

The optimized workflow was applied to the analysis of 45 serum samples of participants with varying degree of liver fibrosis determined independently by clinical evaluation of liver biopsies. Biopsy is the gold standard for evaluation of fibrosis. The Ishak scores were used in our study to classify participants into fibrosis (n = 15) or cirrhosis (n = 15) groups (cirrhosis is an advanced stage of the fibrotic disease). The disease groups are compared to disease-free participants (n = 15).

To test the reproducibility of the LC-MS-MRM measurement, we included a total of 12 QC runs of a pooled sample. The average RSD in 12 QC runs of the 22 core fucosylated glycopeptides is 9.5% (Supplementary Fig. S1A) and 11.7% for the 17 proteotypic peptides (Supplementary Fig. S1B). For majority (~80%) of all the analytes the RSD was below 15%.

The results of quantification of core fucosylation of the 22 glycopeptides, corresponding to 17 proteins, are presented in Table 1. Approximately half of the glycopeptides (12 of 22) increased significantly in core fucosylation compared to the control group. The increase of core fucosylation of some peptides occurred already at the stage of fibrosis while others increased significantly at the cirrhotic stage (see below). Interestingly, decrease in core fucosylation was observed for several glycopeptides as well, in line with previous observations [26]. We did not observe a correlation between the changes in the quantity of the glycoproteins (Supplementary Tables S3) and their core fucosylation. However, the degree of core fucosylation, and its disease-related change, is clearly protein and site specific.

Table 1.

Quantification of changes in core fucosylation of glycopeptides in liver disease.

Protein name	Peptidea	Core fucosylation level in percent (mean ± SD)			p value (ANOVA)
		Control	Fibrosis	Cirrhosis
Ig gamma-2 chain C region	EEQF180NSTFR	91.4 ± 2.75	84.9 ± 13.47	87.5 ± 3.95	0.107
Immuno globulin J chain	E71NISDPTSPLR	16.8 ± 2.6	17.9 ± 4.07	20.3 ± 2.34	0.013
Ig alpha-2 chain C region	TPLTA205NITK	36.2 ± 17.03	30.2 ± 12.97	38 ± 14.7	0.340
Hemopexin	SWPAVG187NCSSALR	9.9 ± 2.41	12.8 ± 3.18	14 ± 5.24	0.016
Complement C3	TVLTPATNHmG85NVTFTIPANR	14.3 ± 10.39	10.4 ± 5.84	11.8 ± 6.6	0.430
Alpha-2-HS-glycoprotein	AALAAFNAQN176NGSNFQLEEISR	15.5 ± 4.57	17.7 ± 4.01	16.8 ± 5.71	0.456
Fibronectin	LDAPTNLQFV1007NETDSTVLVR	98.6 ± 0.64	98.6 ± 0.86	98.4 ± 0.67	0.671
Complement C1q subcomponent subunit A	NPPMGGNWIFDTVITNQEEPYQ146NHSGR	45.3 ± 14.86	49 ± 11.54	49.2 ± 10.63	0.624
Serotransferrin	QQQHLFGS630NVTDCSGNFCLFR	7.7 ± 1.85	10.6 ± 2.17	11.6 ± 4.69	0.004
Plasma protease C1 inhibitor	VLS253NNSDANLELINTWVAK	34 ± 5.46	41.7 ± 5.25	43.8 ± 11.84	0.005
Pigment epithelium-derived, factor	VTQ285NLTLIEESLTSEFIHDIDR	79.9 ± 10.68	76.9 ± 11.52	83.3 ± 6.13	0.224
Alpha-2-macroglobulin	GCVLLSYL55NETVTVSASLESVR	72.3 ± 11.72	66.3 ± 11.72	66.4 ± 10.88	0.271
Clusterin	ML354NTSSLLEQLNEQFNWVSR	5.5 ± 0.95	6.7 ± 1.57	8.6 ± 2.4	0.000
Alpha-1-antitrypsin	QLAHQS70NSTNIFFSPVSIATAFAMLSLGTK	27.9 ± 15.05	27.3 ± 11.24	29.7 ± 11.71	0.875
	ADTHDEILEGLNF107NLTEIPEAQIHEGFQELLR	6.9 ± 1.95	9.9 ± 2.36	11.4 ± 4.48	0.001
Lumican	AFE88NVTDLQWLILDHNLLENSK	91.8 ± 2.75	90.3 ± 10.43	94.4 ± 0.89	0.234
	LHINHN127NLTESVGPLPK	54.5 ± 3.9	58.1 ± 4.84	58.9 ± 4.06	0.016
Vitronectin	169NGSLFAFR	21.7 ± 7.96	14.6 ± 7.92	13.8 ± 7.82	0.017
	N86NATVHEQVGGPSLTSDLQAQSK	5.6 ± 3.15	12.2 ± 6.15	12.6 ± 6.2	0.001
Ceruloplasmin	E397NLTAPGSDSAVFFEQGTTR	9.6 ± 2.34	14.4 ± 2.94	14.9 ± 4.74	0.002
	EHEGAIYPD138NTTDFQR	26 ± 3.86	30.3 ± 5.14	32.4 ± 8.91	0.026
	ELHHLQEQ762NVSNAFLDK	2.1 ± 0.42	2.5 ± 0.38	3.2 ± 1.53	0.009

^aModified glycosylation site is shown in italic.

3.2.1. Increased core fucosylation at the stage of liver fibrosis

Fibrosis of the liver reflects chronic damage associated with the accumulation of extracellular matrix proteins [46,47]. Aberrant glycosylation of these liver proteins has been observed and core fucosylation, in particular, was described in several studies [13–16]. However, quantification of the changes in core fucosylation of the proteins at different stages of the disease progression remains virtually unreported [26]. To the best of our knowledge, multiplex LC-MS-MRM evaluation of core fucosylation of fibrotic liver disease was not previously described. Several studies report increased core fucosylation in the context of hepatocellular carcinoma (HCC), an end-stage complication of the fibrotic liver damage [13–15]. It was, therefore, somewhat surprising that we find the changes in core fucosylation of some proteins already at the stage of liver fibrosis.

We observe increased core fucosylation of 5 glycopeptides at the stage of liver fibrosis (N630 of serotransferrin, N107 of alpha-1-anti-trypsin, N253 of plasma protease C1 inhibitor, N397 of ceruloplasmin, and N86 of vitronectin). In the case of serotransferrin, we observe approximately 8% core fucosylation of the QQQHLFGS⁶³⁰NVT-DCSGNFCLFR glycopeptide in serum of disease-free controls, in agreement with previous reports [48–50]. However, previous studies compared HCC to disease-free controls [15,45,48] while, in our study, core fucosylation increased to 11% in the fibrotic patients (significant 1.4-fold increase) (Fig. 1A). This elevation occurs therefore at a stage of the disease much earlier than the HCC and could be potentially a risk factor for HCC development. The degree of core fucosylation remains elevated but does not further increase in the cirrhotic patients and we do not see differences in the protein abundance of serotransferrin along the progression of liver damage (Supplementary Tables S3).

Fig. 1.

Increased core fucosylation of QQQHLFGS⁶³⁰NVTDCSGNFCLFR of serotransferrin (A) and ADTHDEILEGLNF¹⁰⁷NLTEIPEAQIHEGFQELLR of alpha-1-antitrypsin (B) at the liver fibrosis stage (s, significant at p < 0.05; Ctrl, control; FIB, fibrosis; CIR, cirrhosis).

Similarly, several studies reported core fucosylation of alpha-1-antitrypsin in HCC [15,51–54] but we observe a significant 1.4-fold increase in core fucosylation at the ADTHDEILEGLNF¹⁰⁷NLTEIPEAQIHEGFQELLR peptide in fibrosis and further 1.7-fold increase in cirrhosis compared to the disease-free controls (Fig. 1B). The increase occurs at an earlier stage of the disease progression than previously reported [54], further suggesting that it is indispensable to develop and apply an analytical method of more quantitative precision for the earlier monitoring of core fucosylation level for the better evaluation of liver disease progression. Besides the increased core fucosylation at N107, the protein abundance of alpha-1-antitrypsin also increases at the fibrotic stage (Supplementary Tables S3 and Supplementary Fig. S2A).

In the case of ceruloplasmin, we observe a significant 1.5-fold increase in core fucosylation of E³⁹⁷NLTAPGSDSAVFFEQGTTR at the stage of liver fibrosis; however, core fucosylation of two additional glycopeptides of the same protein increases only in cirrhotic samples (see below). This disease and site-specific change is further pronounced in the case of vitronectin which is also discussed in a later section. The 1.2-fold increase in core fucosylation of the VLS²⁵³NNSDANLELINTWVAK glycopeptide of plasma protease C1 inhibitor in liver fibrosis (or other stage of liver disease) is not to our knowledge previously reported (Table 1).

3.2.2. Increased core fucosylation at the stage of liver cirrhosis

Cirrhosis represents an advanced stage of liver fibrosis and a major risk factor for the development of HCC [55]; several studies of cell and animal models reported that core fucosylation of proteins is a key contributor in the development of HCC [16,20,47]. In our study, two glycopeptides of ceruloplasmin, ELHHLQEQ⁷⁶²NVSNAFLDK (Fig. 2A) and EHEGAIYPD¹³⁸NTTDFQR (Table 1), have significantly up-regulated core fucosylation at the cirrhosis stage, in contrast to the E³⁹⁷NLTAPGSDSAVFFEQGTTR glycopeptide discussed above. This suggests that site-specific changes in glycosylation should have better prediction accuracy at progressive stages of liver disease progression than the analysis of detached glycans which averages across multiple sites and/or proteins.

Fig. 2.

Increased core fucosylation of ELHHLQEQ⁷⁶²NVSNAFLDK of ceruloplasmin (A) and ML³⁵⁴NTSSLLEQLNEQFNWVSR of clusterin (B) significant at the stage of liver cirrhosis. Representative LC-MRM-MS traces of the transition (920.1/1179.6; [M+3H]³⁺) of ML³⁵⁴N(HexNAcFuc)TSSLLEQLNEQFNWVSR (C) and the transition (871.4/1179.6; [M+3H]³⁺) of ML³⁵⁴N(HexNAc)TSSLLEQLNEQFNWVSR (D) of clusterin (s, significant at p < 0.05; ns, non-significant; Ctrl, control; FIB, fibrosis; CIR, cirrhosis).

Clusterin, a well characterized and constitutively secreted extra-cellular chaperone, is also fucosylated [15,26,56,57]. We find a 1.6-fold increase in core fucosylation at ML³⁵⁴NTSSLLEQLNEQFNWVSR of clusterin in liver cirrhosis (Fig. 2B). Fig. 2 shows representative transition of the core fucosylated form (920.1/1179.6; +3 charge; Fig. 2C) and the transition for the HexNAc form (871.4/1179.6; +3 charge; Fig. 2D). The comparison of the retention times on the RP column shows a constant shift of approximately 70 s between the core-fucosylated and non-fucosylated glycoforms of the peptide, which is an expected influence of the addition of core-fucose and thus lower hydrophobicity. Of note, these two glycoforms nearly co-elute for all the other peptides in our study but there is a consistently slightly shorter retention time of the fucosylated glycopeptide (Supplementary Table S1).

The N-glycan profile of hemopexin was suggested as a biomarker for the detection of HCC in patients with cirrhosis [58] and a trend toward increased fucosylation in liver disease was observed in our previously published study [59]. We confirm our observation and show that the core fucosylation of SWPAVG¹⁸⁷NCSSALR glycopeptide of hemopexin increases in cirrhosis. And the discrepancy from ref. [58] may be due to differences in the population examined or differences in the methods. At the same time, the observed decrease in the abundance of hemopexin (Supplementary Fig. S2B) shows that the trends in the changes in protein abundance and core fucosylation are independent. Similar trend (increased core fucosylation and decreased protein abundance in cirrhosis) was observed for the E⁷¹NISDPTSPLR glycopeptide of the immunoglobulin J chain (Supplementary Table S3 and Fig. S2C). Quantification of the core fucosylation provides therefore independent information and may improve our understanding of the liver disease progression. Our results further demonstrate that the targeted quantification of site-specific core fucosylation is feasible even in samples of unfractionated serum without glycopeptide enrichment and that the LC MS-MRM workflow simplifies the previously reported workflows and will be easier to implement in larger studies.

3.2.3. Site-specific core fucosylation of serum proteins

Site-specific changes in core fucosylation are observed for several proteins with more than one core fucosylated glycopeptide. For example, LHINHN¹²⁷NLTESVGPLPK of lumican has increased core fucosylation in cirrhosis but no change in core fucosylation is observed on the other glycopeptide AFE⁸⁸NVTDLQWLILDHNLLENSK (Table 1). The degree of core fucosylation of three glycopeptides of ceruloplasmin in disease-free controls ranges from 2% to 26% (Table 1). Although all three sites (N138, N397, N762) of ceruloplasmin increase in core fucosylation but only one of the peptides increases at the stage of liver fibrosis (Table 1). Opposite change in core fucosylation is observed on two glycopeptides of vitronectin (Fig. 3), a candidate biomarker for the detection of HCC in patients with chronic liver disease [60,61]. Specifically, core fucosylation progressively increases in fibrosis and cirrhosis for the N⁸⁶NATVHEQVGGPS-LTSDLQAQSK glycopeptide (Fig. 3A) but decreases for the ¹⁶⁹NGSLFAFR glycopeptide in cirrhosis (Fig. 3B). We fail to detect reliably a third core fucosylation site of vitronectin observed in previous studies of immunodepleted and/or enriched samples of HCC patients [22,26,62,63]. This might be related to higher degree of branching on one of the sites or the presence of bisected glycoforms [64,65] that were reported to down-regulate core fucosylation [66,67]. It could also be related to the selective retention of specific vitronectin glycoforms in the serum of patients during enrichment [62,63]; we do not attempt to further analyze reasons for the observation. Moreover, no change in the abundance of vitronectin is observed (Supplementary Table S3), in agreement with previous studies [63]. In addition, we find a tendency toward a decrease in core fucosylation of GCVLLSYL⁵⁵NETVTVSASLESVR glycopeptide of alpha-2-macroglobin and EEQF¹⁸⁰NSTFR glycopeptide of Ig gamma-2 chain C region in cirrhosis (Supplementary Fig. S3A and C) although a significant increase in the abundance of each protein is observed (Supplementary Fig. S3B and D). Collectively, the observed differences in core fucosylation at specific glycopeptides clearly document the differential regulation of core fucosylation of individual proteins.

Fig. 3.

Site-specific core fucosylation of N⁸⁶NATVHEQVGGPSLTSDLQAQSK (A) and ¹⁶⁹NGSLFAFR (B) of vitronectin (s, significant at p < 0.05; ns, non-significant; Ctrl, control; FIB, fibrosis; CIR, cirrhosis).

4. Conclusion

In this study, we simplified the detection of core fucosylation of N-glycoproteins and demonstrate the feasibility of direct quantification of these glycoforms without any fractionation of serum or enrichment. The multiplex LC-MS-MRM workflow takes advantage of simplification of the N-glycans with endoglycosidases and the advantageous chromatographic/mass spectrometric properties of the truncated glycopeptides. We demonstrate applicability of the method to clinical samples in a study of liver fibrosis intended to improve our understanding of the elevation in core fucosylation of serum proteins at progressive stages of liver disease. The workflow achieves an average RSD of 9.5% and documents significantly elevated core fucosylation of multiple glycoproteins already at the stage of liver fibrosis which suggests involvement of this pathway at the earliest stages of the disease process. We document that the changes in core fucosylation occur in a protein- and site-specific manner; glycopeptides of some proteins even decrease in core fucosylation against the general increasing trend at successive stages of the disease progression. The increased core fucosylation in fibrotic liver disease needs to be taken into account in the studies of HCC. Although we demonstrate that our LC-MS-MRM work-flow is suitable for the quantification of core fucosylated glycopeptides directly in patient serum, the results do not establish clinical utility of such measurements. Further studies are needed to evaluate sensitivity and specificity of the assays for the detection of liver disease at progressing stages of development. In summary, we describe a simplified LC-MS-MRM workflow which has the potential to improve non-invasive serologic monitoring of liver disease by the quantification of site-specific changes in core fucosylation of proteins.