Tumour mutations in long noncoding RNAs enhance cell fitness

Abstract

Long noncoding RNAs (lncRNAs) can act as tumour suppressor or oncogenes to contrast/promote tumour cell proliferation via RNA-dependent mechanisms. Recently, genome sequencing has identified elevated densities of tumour somatic single nucleotide variants (SNVs) in lncRNA genes. However, this has been attributed to phenotypically-neutral “passenger” processes, and the existence of positively-selected fitness-altering “driver” SNVs acting via lncRNAs has not been addressed. We developed and used ExInAtor2, an improved driver-discovery pipeline, to map pancancer and cancer-specific mutated lncRNAs across an extensive cohort of 2583 primary and 3527 metastatic tumours. The 54 resulting lncRNAs are mostly linked to cancer for the first time. Their significance is supported by a range of clinical and genomic evidence, and display oncogenic potential when experimentally expressed in matched tumour models. Our results revealed a striking SNV hotspot in the iconic NEAT1 oncogene, which was ascribed by previous studies to passenger processes. To directly evaluate the functional significance of NEAT1 SNVs, we used in cellulo mutagenesis to introduce tumour-like mutations in the gene and observed a consequent increase in cell proliferation in both transformed and normal backgrounds. Mechanistic analyses revealed that SNVs alter NEAT1 ribonucleoprotein assembly and boost subnuclear paraspeckles. This is the first experimental evidence that mutated lncRNAs can contribute to the pathological fitness of tumour cells.

CADTAD: CAncer Driver Topologically Associated Domains identify oncogenic and tumor suppressive lncRNAs

“…In addition, we observed almost half of the CDLs identified by a recent publication 37 located in CDTs (Fig. 1I).…”

“…This is achieved by estimating the cumulative functional impact bias of somatic mutations within tumors, using signals of positive selection as a guiding principle. Furthermore, ExInAtor1/2 37,49 provides a framework to identify driver lncRNAs by discerning driver mutations within them. This identification is rooted in the detection of signals of positive selection that act on somatic mutations, thereby considering the mutational burden and functional impact.…”

“…The current methods for discovering CDLs have primarily remained at the level of CRLs, lacking the specificity to differentiate their precise roles in either promoting or inhibiting cancer progression. Second, the identification of driver lncRNAs based on genomic characteristics has yielded relatively low numbers 37,48,49 . To address this, we advocate approaching CDLs from a distinct perspective, one that is different from the paradigm of cancer driver protein-coding genes.…”

“…RefLnc 58 was used to supplement the TCGA (PRAD, STAD, LUSC, and LUAD) and GTEx projects with combined FPKM data to obtain the expression of 26494 lncRNAs and 19490 coding genes in a single cancer type. The 17 driver lncRNAs identified by ExInAtor2 were from a publication 37 .…”