Introduction
Protein expression and post-translational modifications, including glycosylation, play a pivotal role in the onset and progression of colorectal cancer. Protein glycosylation, particularly O-linked glycosylation, contributes to tumor cell signaling, adhesion, and immune evasion. Abnormal proteomic and glycoproteomic patterns have been shown to correlate with cancer stage and aggressiveness. In this study, we combined proteomic and O-glycoproteomic analyses with machine learning approaches to identify molecular signatures and develop predictive models for risk stratification in patients with colorectal cancer.
Methods
Serum, tumor tissue, and matched healthy margin samples were processed using an automated protocol implemented on the Andrew Alliance Andrew Robotic System. Proteins were digested and analyzed to obtain both global proteomic and O-glycoproteomic profiles. Separation of peptides and glycopeptides was achieved using the Waters Premier UHPLC system. Mass spectrometric analyses were performed on a cyclic ion mobility quadrupole time-of-flight mass spectrometer — the SELECT SERIES Cyclic IMS (Waters). Data were processed using dedicated bioinformatics pipelines integrating proteomic and O-glycoproteomic datasets.
Results
Comprehensive proteomic and O-glycoproteomic profiling revealed distinct molecular patterns between colorectal cancer stages I–IV. Significant differences were observed in protein levels and site-specific O-glycosylation profiles across tumor and non-tumor tissues, suggesting stage-dependent remodeling of the proteome and glycoproteome.
Conclusions
Integrative proteomic and O-glycoproteomic analyses offer novel insights into the molecular mechanisms of colorectal cancer progression. These approaches hold great potential for improving diagnostic accuracy, patient stratification, and the identification of new therapeutic targets in translational oncology.