Circular ANRIL isoforms switch from repressors to activators of p15/CDKN2B expression during RAF1 oncogene-induced senescence

Abstract

The ANRIL ncRNA is produced as an antisense RNA to the p15-encoding gene from the INK4 locus, which encodes three tumor suppressor proteins required for cell cycle arrest, p15 INK4b , p16 INK4a and p14 ARF . In proliferative cells, ANRIL prevents cellular senescence by repressing the expression of INK4 genes through Polycomb protein recruitment. Decreased expression of ANRIL in senescence is thought to relieve this repression. Here, we observe in various models of oncogene-induced senescence that senescence induction is accompanied by an increase in ANRIL expression, including the increase of circular ANRIL species. Circular ANRIL species repress p15 gene expression in proliferative cells as previously described for ANRIL. However, in RAF1 oncogeneinduced senescent cells, circular ANRIL isoforms are required for full induction of p15, p16 and p14 ARF gene expression. They interact with Polycomb group proteins and regulate H3K27me3 at the p16 gene promoter. Hence, we propose a mechanism relying on changes in the ratio between circular ANRIL and Polycomb proteins by which circular ANRIL species switch from being repressors of p15 expression to activators of INK4 gene expression during RAF1-induced senescence. Our data highlight the functional importance of the ratio between ncRNAs and their interacting effectors in the control of cell fate progression.

High-throughput analysis of ANRIL circRNA isoforms in human pancreatic islets

“…Previous studies in immune, cancer, and epithelial cells have identified a wide range of circANRIL isoforms that are expressed in different conditions (17, 18, 20, 30). However, such an analysis has not been conducted in pancreatic islets, despite evidence that circular and linear ANRIL expression may be implicated in diabetes phenotypes.…”

“…Notably, the abundance of circANRIL isoforms were more correlated with each other than with linANRIL or other mRNAs transcribed from the locus, suggesting independent, non-transcriptional regulation of these distinct ANRIL molecules. Previous studies have also seen that overall circANRIL abundance may be more correlated with CDKN2A and CDKN2B isoforms and perhaps even regulate the abundance of these mRNAs (30). Together, these results highlight the need for a greater understanding of circANRIL regulation across different cell types.…”

“…Previous studies seeking to identify circANRIL molecules have mostly used Sanger sequencing after divergent primer PCR (17, 18, 20, 30, 47), which limits the detection of isoforms due to primer selection and placement. Our unbiased high-throughput sequencing approach allowed us to identify multiple distinct isoforms for circANRIL, with extensive alternative splicing leading to varied internal exon structures.…”

“…Since these molecules have no 5’ cap or 3’ polyA tail, they are not believed to be translated and cannot be targeted by exonucleolytic degradation. Thus, circANRIL isoforms are thought to have greater stability than linANRIL species (17, 20, 30) and, consistently, have been seen at greater levels than linANRIL in immune cells (17, 20). The handful of studies that have characterized ANRIL exons involved in back-splicing events have not identified any cis-elements that may promote back-splicing.…”

Emerging Roles of LncRNAs in the EZH2-regulated Oncogenic Network

“…In a study of oncogene-induced senescence, Muniz et al discovered that an isoform of the circular ANRIL could bind to Polycomb proteins and reduce EZH2 occupation on the p15 and p16 promoters, leading to declined H3K27me3. As a result, the expression of ANRIL or its relative levels versus EZH2 determined the expression of these genes, as well as cell senescence [112]. Noteworthily, many lncRNAs are downstream effectors of EZH2, and over 20% of lncRNAs are regulated by PRC2 (Figure 2A) [113].…”