Abstract
Aberrant changes in site-specific core fucosylation (CF) of serum proteins contribute to cancer development and progression, which enables them as potential diagnostic markers of tumors. An optimized data-dependent acquisition (DDA) workflow involving isobaric tags for relative and absolute quantitation-labeling and enrichment of CF peptides by lens culinaris lectin was applied to identify CF of serum proteins in a test set of patients with nonalcoholic steatohepatitis (NASH)-related cirrhosis (N = 16) and hepatocellular carcinoma (HCC, N = 17), respectively. A total of 624 CF peptides from 343 proteins, with 683 CF sites, were identified in our DDA-mass spectrometry (MS) analysis. Subsequently, 19 candidate CF peptide markers were evaluated by a target parallel reaction-monitoring-MS workflow in a validation set of 58 patients, including NASH-related cirrhosis (N = 29), early-stage HCC (N = 21), and late-stage HCC (N = 8). Significant changes (p < 0.01) were observed in four CF peptides between cirrhosis and HCC, where peptide LGSFEGLVn160LTFIHLQHNR from LUM in combination with AFP showed the best diagnostic performance in discriminating HCC from cirrhosis, with an area under curve (AUC) of 0.855 compared to AFP only (AUC = 0.717). This peptide in combination with AFP also significantly improved diagnostic performance in distinguishing early HCC from cirrhosis, with an AUC of 0.839 compared to AFP only (AUC = 0.689). Validation of this novel promising biomarker panel in larger cohorts should be performed.