|dc.description.abstract||Protein glycosylation is one of the most common co- and post-translational modifications occurring on proteins. It impacts vital biological roles such as cell-cell interactions, recognition events, host and pathogen interactions, adaptive and innate immunity among many others. In diseases such as cancer, changes in protein glycosylation are ubiquitously present and considered a key feature of malignancy. The ability to determine and quantitate these changes has long been a focus of glycomics and glycoproteomics to understand the role of protein glycosylation in health and disease.
A vast majority of food and drug administration approved protein-based cancer markers are glycoproteins. To date, however, the potential of using their specific glycosylation signatures to improve cancer diagnostics sensitivity and selectivity is under-utilised. In my thesis I performed to date the most comprehensive glycomic and glycoproteomic characterisation of carcinoembryonic antigen (CEA), an important gastrointestinal marker, to evaluate if its glycosylation can provide diagnostically relevant information with the potential to improve current diagnostics. I used porous-graphitised-carbon chromatography electrospray ionisation and tandem mass spectrometry (PGC-nanoLC-ESI-MS/MS) glycomics technologies to uncover potential novel targets for gastric cancer (GC) in a GC cell line model. The same approach was employed to investigate the glycomic profile of the tumour marker epithelial cadherin (E-cadherin) that is involved in the pathogenesis of gastric cancer. I was also teaming up with bioinformaticians to develop a web-based tool to aid the manual identification of intact glycopeptides, which I also applied in the site-specific glycosylation analysis of CEA.
For the comprehensive glycomics and glycoproteomics analyses CEA purified from four different origins was used. It was obtained from human colon cancer (cell line and tissue), tissue from liver metastasis of colon cancer and ascites fluid. These different CEAs exhibited distinct N-glycosylation differences such as N-glycan branching, degree of sialylation, antenna fucosylation and the level of bisecting N-glycans. CEA was also verified to be a carrier of sialyl Lewis x as validated by specific exoglycosidases digestions. More importantly, an unusual glyco-epitope, a 1-4 galactosidase resistant hexose attached to the bisecting GlcNAc, was identified to be present in high abundance on colon cancer derived CEA. A site-specific glycosylation analysis of CEA was achieved by digestion with pronase and analysis of the resulting glycopeptides in a dual LC system combining reverse-phase and PGC chromatography within a single analysis using electrospray ionisation and tandem mass spectrometry detection. 27 out of the 28 predicted N-glycosylation sites were identified to be glycosylated and a novel N-glycosylation site located within a non-canonical 76N-R-Q78 sequence motif was also identified.
In a collaborative project I was using PGC-nanoLC-ESI-MS/MS glycomics to investigate the role of sialylation in GC. MKN45 GC cells transfected with ST3GAL4 showed a significant increase of 2-3 sialylation on N-glycans while there was a substantial reduction of N-glycan 2-6 sialylation. A similar trend was also observed for core 2 O-GalNAc glycans. Moreover, lower levels of bisected N-glycans were also detected. Analysis of the sialoproteome of these cells identified RON receptor kinase and CEA among the glycoproteins identified to exhibit increased levels of sialylation. Glycomic characterisation of immunoprecipitated CEA derived from ST3GAL4 overexpressing MKN45 GC cells also confirmed higher levels of 2-3 sialylation. Further investigations confirmed that increased 2-3 sialylation levels on CEA were a clinicopathological feature of gastric tumours.
I was also involved in another collaborative project investigating the role of site-specific N-glycosylation for E-cadherin function. Various different E-cadherin glycoforms were produced by site-specific mutagenesis and expression in a GC cell line. Deletion of single sites affected the E-cadherin specific N-glycome and identified that in particular N-glycosylation on site Asn554 is relevant for E-cadherin function.
In collaboration with bioinformaticians from the Swiss Institute of Bioinformatics a web-based tool called PepSweetener was developed to help the manual identification of intact glycopeptides. This tool was designed to generate a matrix containing theoretical glycopeptides corresponding to precursor masses that fall within a specified error range. The resulting matrix is an interactive heatmap that allows the user to identify the most probable glycopeptide by matching the peptide sequence with the theoretical fragmentation diagram of the selected peptide sequence speeding up the analysis.||en_US