Expanding horizons: new roles for non-canonical RNA-binding proteins in cancer

Abstract

Cancer development involves the stepwise accumulation of genetic lesions that overcome the normal regulatory pathways that prevent unconstrained cell division and tissue growth. Identification of the genetic changes that cause cancer has long been the subject of intensive study, leading to the identification of several RNA-binding proteins (RBPs) linked to cancer. Cross-reference of the complement of RBPs recently identified by RNA interactome capture with cancer-associated genes and biological processes led to the identification of a set of 411 proteins with potential implications in cancer biology. These involve a broad spectrum of cellular processes including response to stress, metabolism and cell adhesion. Future studies should aim to understand these proteins and their connection to cancer from an RNA-centred perspective, holding the promise of new mechanistic understanding of cancer formation and novel approaches to diagnosis and treatment.

Figure 1

RNA-binding proteins linked to cancer. (a) STRING [61] network showing connections between RNA-binding proteins (RBPs) with links to cancer based on annotations in COSMIC, OMIM, and GO (see main text). Main network hubs are highlighted. (b) Sunburst graph showing breakdown of cancer-linked RBPs by RBD classification (i.e. classical or non-canonical) and source of cancer-link information. (c) Word cloud of domains present in unorthodox cancer-linked RBPs. Size indicates relative prevalence.

Figure 2

Examples of unorthodox cancer-linked RBPs. Architecture of (a) RNA-binding HSP90 proteins (b) IDH proteins 1 and 2 (c) the three ERM proteins. Lollipops indicate the cancer-associated mutations available in the ICGC data portal [62], black boxes represent Pfam-annotated [63] protein domains, red and orange boxes map the high-confidence and candidate RNA-binding sites reported by RBDmap [20^••], respectively, and green boxes indicate regions which are predicted to be intrinsically disordered (IUPred score > 0.4) [64]. Frequently mutated residues in IDH1 and IDH2 proteins with established links to several cancers (see text for further information), are denoted by the number of reported mutations in parentheses.