Abstract:
Aberrant glycosylation is a defining feature of cancer, influencing tumor initiation, progression, and metastasis. This thesis integrates three dimensions of glycobiology—glycosyltransferase (GT) expression, glycosylated RNAs (glycoRNAs), and E-selectin-binding glycoconjugates—to establish new diagnostic and mechanistic frameworks. A 71-gene cancer-pattern GT (CPGT) panel enabled machine-learning models that accurately classified and predicted cancer subtypes and prognosis, outperforming conventional assays. GlycoRNA profiling in acute myeloid leukemia revealed sialoglycoRNAs enriched in both large and small extracellular vesicles (EVs), linked to glycan biosynthesis, signaling, and neuronal regulation, with H1 RNA emerging as a key glycoRNA. Proteomic mapping further uncovered conserved glycoRNA-binding proteins across trafficking and mitochondrial compartments. Finally, dissection of E-selectin-binding ligands in breast cancer uncovered subtype-specific repertoires spanning glycoproteins, glycolipids, and glycoRNAs. Collectively, these studies reveal how glycosylation—encoded by GTs, embedded in RNAs, and displayed on cell surfaces—drives cancer heterogeneity and offers new avenues for biomarker and therapeutic discovery.

Bio:
Jing Kai is a PhD candidate in Prof. Jasmeen Merzaban's lab, where her research centers on glycobiology and cancer. Her work combines experimental and computational approaches to explore glycosyltransferase gene signatures for cancer classification and prognosis, and to uncover the biology of glycoRNAs and the functional roles of E-selectin ligands in tumor progression and extracellular vesicle communication.