A4GALT-related glycosphingolipids play a pivotal role in the reversible transition of mesenchymal and epithelial ovarian cancer cells and hence are important regulators of metastasis

Introduction:

The transition between epithelial and mesenchymal states (EMT) comprises cellular and molecular processes essential for local tumor growth and dissemination. We investigate the function of glycosphingolipids (GSL) in EMT in ovarian cancer and their role in disease outcome.

Materials and methods:

CRISPR-Cas9-technology was employed to generate A4GALT-knockout sublines from several parental ovarian cancer cell lines. GSL expression profile was determined by flow cytometry. Gene and protein expression were determined by RT-qPCR and Western blotting, respectively. Publicly available transcriptomic data sets were used to compute overall survival and relapse-free survival.

Results:

Ovarian cancer cells with epithelial features have elevated expression of genes encoding glycosyltransferases (e.g. A4GALT) responsible for globoside biosynthesis and this elevated expression was associated with better outcome (OS, RFS) in ovarian cancer patients. Deletion of A4GALT and subsequent depletion of globosides induced EMT-typical changes: switch from cobble-stone to fibroblast-like morphology; loss of E-cadherin expression as consequence of epigenetic silencing of CDH1; elevated capacities to migrate and disseminate (zebrafish-model), anchorage-independent (anoikis-resistance) proliferation; chemoresistance; adaptation of cancer stem cell-like features. Intriguingly, accurate E-cadherin-mediated cell-cell adhesion required both functional A4GALT and globosides.

Conclusion:

We propose a model in which A4GALT function and expression of globosides are required and accurate expression/localization of E-cadherin are essential for the epithelial cell state and in which, conversely, A4GALT deletion-imposed loss of globosides and concomitant epigenetic silencing of E-cadherin expression induces EMT associated with chemoresistance and acquisition of cancer stem cell-like features. This model reveals potential therapeutic targets to prevent cancer dissemination.