Exogenous expression of the glycosyltransferase LARGE1 restores α-dystroglycan matriglycan and laminin binding in rhabdomyosarcoma.

Abstract

Background: α-Dystroglycan is the highly glycosylated component of the dystrophin-glycoprotein complex (DGC) that binds with high-affinity to extracellular matrix (ECM) proteins containing laminin-G-like (LG) domains via a unique heteropolysaccharide [-GlcA-beta1,3-Xyl-alpha1,3-]_n called matriglycan. Changes in expression of components of the DGC or in the O-glycosylation of α-dystroglycan result in muscular dystrophy but are also observed in certain cancers. In mice, the loss of either of two DGC proteins, dystrophin or α-sarcoglycan, is associated with a high incidence of rhabdomyosarcoma (RMS). In addition, glycosylation of α-dystroglycan is aberrant in a small cohort of human patients with RMS. Since both the glycosylation of α-dystroglycan and its function as an ECM receptor require over 18 post-translational processing enzymes, we hypothesized that understanding its role in the pathogenesis of RMS requires a complete analysis of the expression of dystroglycan-modifying enzymes and the characterization of α-dystroglycan glycosylation in the context of RMS.

Methods: A series of cell lines and biopsy samples from human and mouse RMS were analyzed for the glycosylation status of α-dystroglycan and for expression of the genes encoding the responsible enzymes, in particular those required for the addition of matriglycan. Furthermore, the glycosyltransferase LARGE1 was ectopically expressed in RMS cells to determine its effects on matriglycan modifications and the ability of α-dystroglycan to function as a laminin receptor.

Results: Immunohistochemistry and immunoblotting of a collection of primary RMS tumors show that although α-dystroglycan is consistently expressed and glycosylated in these tumors, α-dystroglycan lacks matriglycan and the ability to bind laminin. Similarly, in a series of cell lines derived from human and mouse RMS, α-dystroglycan lacks matriglycan modification and the ability to bind laminin. RNAseq data from RMS cell lines was analyzed for expression of the genes known to be involved in α-dystroglycan glycosylation, which revealed that, for most cell lines, the lack of matriglycan can be attributed to the downregulation of the dystroglycan-modifying enzyme LARGE1. Ectopic expression of LARGE1 in these cell cultures restored matriglycan to levels comparable to those in muscle and restored high-affinity laminin binding to α-dystroglycan.

Conclusions: Collectively, our findings demonstrate that a lack of matriglycan on α-dystroglycan is a common feature in RMS due to the downregulation of LARGE1, and that ectopic expression of LARGE1 can restore matriglycan modifications and the ability of α-dystroglycan to function as an ECM receptor.